Advanced Collaborative Distance Learning Systems for young students: Design issues and current trends on new cognitive and meta-cognitive tools
Angelique Dimitracopoulou & Argyro Petrou Learning Technology and Educational Engineering Laboratory Department of Sciences of Pre-School Education and of the Educational Design, University of the Aegean, 1, Av. Democratias, GR 85100, Greece Tel: +22410-99127, Fax: +22410 99175 Email:
[email protected]
-1-
Advanced Collaborative Distance Learning Systems for young students: Design issues and current trends on new cognitive and meta-cognitive tools
ABSTRACT This paper reviews recent applications concerning technology-based learning systems that promote collaboration among students by distance. The research community in order to support learning as well as collaboration, has designed systems differentiating themselves from common web-based tools (simply enabling collaborative activities), and constitute new cognitive and metacognitive tools. The paper, firstly, presents some distinctions of collaborative learning systems from other networked systems, based on assumptions on collaborative learning processes and possibilities of implementation in current educational settings. Secondly, by presenting some cross section systems proceeds to an analysis and categorization of main tools and functionalities that characterizing them. Thirdly, it analyses and discusses extensively the current trade-offs of collaborative systems design. Finally, it attempts a synthesis of necessary tools by a proposed framework and concludes by presenting the main axes of the research agenda concerning design of systems and underlying tools addressed to young students of primary and secondary education.
INTRODUCTION During the last decade, the second part of which there has been an intense research interest, a new and significant community of researchers and practitioners has emerged focusing on collaborative computer-supported learning systems and collaborative learning. The community and the underlying research is usually expressed through the ‘single’ CSCL, that stands for “Computer-Supported Collaborative Learning” or recently, for the wider “Computer Support for Collaborative Learning” (Koschmann, 1999). The
-2-
impressive evolution of this new research domain and of development orientation is due to two burgeoning advancement of two parallel fields of research: (a)
The evolution and a shift of learning theories: The importance of social process of learning, including the potential utility of collaborative learning, is actually well established. New theories of learning have emerged and developed during the last decade pointing out the social dimension of learning process. Theories such as Distributed cognition theory (Hutchins, 1991; Salomon, 1995; Pea, 1995; Hollan, Hutchins & Kirsh, 2000), Activity Theory (Engestrom, 1987; Bodker; 1991; Kuutti, 1996; Jonassen, 2000), Situated Learning (Resnick et all., 1991; Clancey, 1997; Greeno et all, 1993) had a significant impact, while at the same time, the new learning situations that technologies have made possible have provided a new field of research on one hand, and on the other the possibility of further development of these theories.
(b)
The advances in current information and communications technology: have provided many new opportunities for various forms of communication, cooperation and collaboration via networks. The acceptance of the Web and the recognition of the importance of distance education in various countries have made advances in CSCL particularly urgent. In fact, broad concerns about the limitations of traditional educational approaches in an increasingly global and technological world underscore the need to realize the potentials of collaborative learning and computer support. This new research field has already produced technologies through a significant
number of technology-supported systems that promote learning through various collaborative settings of users, learning contexts and conditions. These systems are intended to mediate collaboration, cognitive development as well as to make knowledge
-3-
and knowledge-building communities more broadly accessible. At the same time, the field has developed a theoretical understanding of learning as a collaborative process that is socially and culturally situated, producing ways of observing and assessing collaborative learning (with co-presence or by distance) and knowledge building. Since the first congress on CSCL (hold 1995 in Indiana) until today, the genuinely interdisciplinary community, incorporates researchers, designers, or educators, from various disciplines including: education, cognitive, social and educational psychology, didactics, subject matter specialties, computer science, linguistics and semiotics, speech communication, anthropology, sociology and design. The community works on theoretical frameworks, on design of components and artifacts, appropriate architectures and methods of development, methods of significant qualitative or quantitative evaluation of collaborative situations, approaches of implementation in actual educational systems, collaborative learning activities and pedagogical approaches, new roles of various implicated agents (such as students, teachers), etc., aiming at producing tools and systems, developing our understanding of learning processes and trying to find the best ways to implement the approaches and tools in the actual educational systems. The reader may find significant review papers on the evolution of research on collaborative learning, (Dillenbourg et all 1996) or on the meaning of collaborative learning (Dillenbourg 1999), on epistemological foundations of CSCL (Lipponen 2002; Paavola, Lipponen & Hakkarainen 2002), on learning effects and best practices (Lehtinen et all 1998). All of them constitute theoretical reviews and foundational papers. One interesting review oriented to the design that analyses the specific design tools and functionalities (Jermann, Soller & Muehlenbrock, 2001) whose purpose is to support learners during collaborative activities.
-4-
But what are the main design achievements of this research field, what are the current trade-offs? What are the new research directions on design? The exploration of these general questions is the objective of the present paper. Initially, some of the questions that we try to clarify are the following: •
What differentiate the collaborative learning systems from other web-based systems and tools?
•
What are the characteristics of some main systems that were designed mainly for students of secondary and primary education? The first question is a clarification of the field. The second question is answered
through an analysis of some characteristic systems, and lead to the categorization of the main tools and functionalities of collaborative learning systems. Finally, the paper focuses on the following questions: •
What are the main trade-offs actually in the field? What are the various designers’ choices concerning some significant design aspects?
•
What are the main actual design questions that preoccupy researchers and form a research agenda for the next years? For this final question, authors specify a framework of analysis that is derived from
their consideration of the social environment, in a school context. Some main open questions in the current research state are discussed and main axes of the actual and future research agenda are pointed out concerning design of systems and tools addressed to young students of primary and secondary education. Briefly, the paper can be viewed as consisting of four main parts. In the first part, the differentiation of the CSCL systems to various existing web based systems and tools are discussed. Then, the main reasons that educational technology designers may be interested in CSCL systems are outlined. The second part presents briefly some
-5-
characteristic systems and proceeds to an analysis and categorisation of their main tools and functionalities. The third part, discusses the current trade offs on the design. The fourth and final part, make a synthesis by a proposed framework and points out some axes of research agenda for the community in the next years.
WHAT
SPECIFIC
ON
TECHNOLOGY
SUPPORTED
SYSTEMS
FOR
COLLABORATIVE LEARNING? At present, almost any web-based application is labelled as ‘collaborative’. This loose usage, appear when a word becomes fashionable: collaboration appears often as a synonym for a good learning and good educational technology. This situation is accentuated also by the fact that there is not an established way to classify the variety of tools that might be considered collaborative. Furthermore, it is true that almost any technological application, could, in some way, be used by people collaborating, i.e., by people working together on something. Hence, it might be meaningful to make first of all, a distinction between collaborative use of technology and collaborative technology, and it is this second concept that we examine here. Then, it is necessary to differentiate the systems that simply allow collaboration from the systems that support collaboration. And lets examine three categories: (a) the various tools on Internet, (b) the platforms for complete educational programs by distance, (c) the systems for collaborative work. (a) Collaboration can take place through computer networks, those most well known on the Internet, but not without special efforts. Typical Internet chat, e-mail, bulleting board systems, web-based conferencing, or shared data-bases are applications that could equally be utilised for collaborative learning (under a good combination and advanced
-6-
well-designed pedagogical practices). These applications were most initially designed for pedagogical purposes of building collaborative knowledge, and among anything else, they do not organise sufficiently conversations which promote learning. Furthermore, most of the Internet tools and discussion forums available are not robust and simple enough to use in average classrooms, or do not fit in the classroom settings. (b) Another category of systems that could allow collaboration are the systems that aim at providing a global platform for a whole educational program, such as WebCT, Learning Space, CENTRA, FirstClass, etc.,(Mason, 1998; Bratitsis & Dimitracopoulou 2001). These systems employ computing resources to do things that educators have previously done without computers. Accepting as a model the traditional classrooms, they can eventually contribute to improve just a little some aspects of teaching, such as preparation of teaching materials and class management. Most of the early Web-mediated on-line courses were designed to complement conventional methodologies for dissemination of course content, connecting students to various on-line multimedia learning materials. Generally, the implementation of these systems does not require the change of the teaching strategies and schemata. On the contrary, the collaborative systems that we are interested in studying here are all oriented in bringing new teaching and learning methods and experiences that could not be implemented by traditional means. In the context of networked multimedia computing, these methodologies carry the potential of fundamental improvements in the efficiency of flexible on-line teaching and learning processes, though the initial costs and investments are much higher than in the development of traditional curriculum. Concerning the learners themselves, traditional platforms for distance education, dispose tools and functions that facilitate the exchange of information or materials (that is communication), without offering specific tools that could really allow collaboration among them.
-7-
(c) An interesting category of systems are the ones that are usually called groupwares and constitute the product of another scientific field, that of Computer Supported Collaborative
Work (CSCW). A CSCW system is defined as a computer-based network system that supports group work in a common task and provides a shared interface for groups to work with (Ellis et al. 1991). They often constitute powerful systems for collaborative work by distance among various organizations, whose mentality needs to be transformed, so as to accept these new modes of work (Vandenbosch & Ginzberg 1997). Most of the groupware applications (e.g. Lotus Notes , TeamFocus , Teamware ), dispose or support advanced discussion databases, and serve as development platforms on which highly structured databases or workflow applications can be build. The main differences between CSCW and CSCL systems are that (a) CSCW is used mainly in the business setting, CSCL is used in the educational setting; (b) CSCW tends to focus on communication techniques themselves, while CSCL focuses on what is being communicated; (c) the purpose of CSCW is to facilitate group communication and productivity, while the purpose of CSCL is to scaffold or support students in learning together effectively, offering for that purpose specific cognitive and meta-cognitive tools. Despite the above, it is to be noted that CSCW research has put forward many issues about cooperative nature of work in the technology based work context. Some theoretical ideas and tools used in collaborative learning environments were originally been created and elaborated in modern work contexts. Thus, in a sense, “CSCL is the youngest sibling of CSCW research field” (Lipponen, 2002).
Summarising the above paragraphs, we could differentiate systems for collaborative learning, from the various tools on web, the platforms for distance education programs, the systems for collaborative work for adults, by taking into account that CSCL systems:
-8-
Promote learning: the purpose is not just to perform a task (systems for collaborative work do) but promote learning during users activity. Enable collaboration among participants during a specific activity: collaboration is distinguished from simple interaction between participants, from simple exchange of ideas or information or materials that concerns mainly communication. Most of the integrated systems for distance education, offers tools to communicate and exchange but not really to collaborate. Support collaboration: CSCL systems they aren’t restrained to allow collaboration, like most of the group networked system, through specific tools. The analysis that follows of some of the existing collaborative learning systems will reveal the status of tools and the functions that support learning.
WHY ARE WE INTERESTED IN COLLABORATIVE LEARNING SYSTEMS? The research community has tried to explore the potential of significant learning effects of all kinds of collaborative settings, related to time and space dimensions: face-to-face collaboration, synchronous and asynchronous collaboration, through local and wide networks, via specific collaborative systems. The wide interest in technology supported collaborative learning is based, in our consideration, on two complementary reasons: reasons referring to social order but most important cognitive ones. a) Social order reasons, reflecting a dominant work mode of actual and future life: Learn to communicate and collaborate is an important skill of actual life. It seems that one of the basic requirements of future education will be to prepare learners for participation in a networked, information society in which knowledge will be the most critical resource for social and economic development. Collaboration is integral to today’s organisations, which require individuals who can work together to solve complex problems and share
-9-
their own knowledge and expertise with others. Collaborative skills can be learned, and it is therefore essential to provide individuals with appropriate learning opportunities (Constantino-Consalez & Suthers, 2001). It is considered that if students remain in social and intellectual isolation, they may fail to develop and refine those cognitive and interpersonal skills increasingly necessary for social and professional life (Abrami 1996). b) Cognitive reasons: Studies on the achievement effects of cooperative learning have taken place in every major subject, at all grade levels. As a result of these researches there is a growing consensus among researchers about the potential of positive effects of collaborative learning on students’ achievement (Slavin, 1995, 1997), under appropriate conditions concerning the tasks, the specific collaborative settings or the roles of teachers. The conversation, the multiple perspectives and the argumentation that arise in collaborative groups may explain why the members of these groups facilitate greater cognitive development than the same individuals achieve when working alone (Harasime 1997). We could distinguish three main reasons that underlie the learning effect of collaborative settings: motivational, social cohesion related reasons (Lehtinen et all 1999) and cognitive elaboration reasons. -
Motivational reasons considering the reward or goal structures under which students operate. Cooperative incentive structures create a situation in which the only way group members can attain their personal goals is if all the members of the group are successful. In these conditions, group members usually both help their group mates to do whatever helps the group to succeed and to encourage their group-mates to exert maximum efforts.
-
Social cohesion reasons, based on the idea that students help their group-mates learn because they care about the group. The social cohesion perspective emphasizes team
- 10 -
building activities in preparation for cooperative learning, as well as group self evaluation, instead of external incentives and individual accountability. A well known application of this interpretation is Aronson’s Jigsaw approach for learning activities (Aronson, et al., 1978), where students concentrate on different topics in ‘expert groups’ and subsequently share their expertise in groups. The theoretical idea in a Jisgaw method is to create interdependence between members in a way that would increase social cohesion. Therefore, distribution of a task among several agents has fundamental cognitive significance. -
Cognitive reasons: collaborative learning is assumed to be effective because it requires participants to elaborate their cognitive structures in a social context. Miyake (1986) and Hutchins (1995) have argued that social interaction (and interaction with the tools of technological culture) provides new cognitive resources for human cognitive accomplishment. In a shared problem solving process, agents who have partial but different information about the problem in question both appear to improve their understanding through social interaction. Under these circumstances, the contradictions, inconsistencies and limitations of an agent’ explanations become available because they force the agent to perceive his or her conceptualisations from different points of view. Externalisation is an important prerequisite of socially distributed cognitive achievements. Additionally, limited cognitive resources can be overcome by distributing the cognitive load to several agents, each of whom is equipped with a restricted power of cognition. Moreover, social interaction fosters the emergence of a more abstract conception than individual working (Schwartz 1995).
Collaborative Technology offers the kind of potential for learning which is very different from those available in other contexts. Beyond the self-evident benefit, that these systems break down the physical and temporal barriers of schooling by removing time and space
- 11 -
constraints, the most essential is that an important number of empirical research has revealed a long list of the promises and report benefits from technology based collaboration and the underlying tools (for an extended review see Lehtinen, et all 1998). Specifically, the most salient are presented below: •
In the case of asynchronous communication, the delay occurred allows time for reflection in interaction.
•
In the case of synchronous collaboration, the students must externalise and argument on their actions in a common and shared space. This externalisation and argumentation are most precious in concept-centred action domains.
•
Making thinking visible by writing should help students to reflect on their own and other’s ideas and share their expertise. Obviously, a cooperative group does not automatically improve the construction of higher order cognitive skills and complex knowledge structures. In order to increase the possibilities for mutual understanding and task-related social interaction, interaction tools that are adequately related both to the new concepts to be learned and to the previous experience and knowledge of students are needed (Katz & Lesgold, 1993).
•
Shared discourse spaces and distributed interaction can offer multiple perspectives and zones of proximal development for students with varying knowledge and competencies.
•
Specific tools such as database can function as a collective memory for a learning community, storing the history of knowledge construction processes for revisions and future use. Furthermore, some tools that analyse the learning process and/or the collaboration process, could offer a metacognitive support precious for learning.
In general, CSCL environments can offer greater opportunities to share and solicit knowledge. In most situations, externalization, articulation, argumentation, negotiation of
- 12 -
multiple perspectives, are considered being the main mechanisms that can promote collaborative learning (Dillenbourg and Schneider, 1995; Baker, 1996; Veerman, 2000). In every case, it should be noticed, that learning does not occur in every day collaborative situations. Effective collaboration depends on, the age, the task, the group composition as well as the role distribution, the moderator (teacher), and evidently, the available tools, and the whole instructional design (Dillenbourg, 1996, 2002).
WHAT ARE THE MAIN FEATURES OF CHARACTERISTIC COLLABORATIVE LEARNING SYSTEMS? The design of collaborative learning environments is concerned with how best to support effective collaborative learning. There are many possibilities. There are systems that are content specific, (for instance for collaborative resolution of specific problems in physics), while others constitute context-free multipurpose environments. In order to present in an exemplary purpose some of collaborative learning systems, we have chosen, ten systems considering them as most characteristics and known in the current community: DIALAB, CoVis, BELVEDERE, Knowledge Forum, COLER, C-CHENE, BetterBlether, Group Leader Tutor, DEGREE and COMET (see Table 1). About half of these systems were designed for students in primary and secondary education, while the other half is composed by systems that were evaluated with students of higher education, but we consider that they could be also used in specific courses of higher level of secondary education. Each of the selected systems has to present a specific/particular tool or feature that has influenced the consecutive design of other systems in its era. The brief description of these systems presents also the context of use and the age of pupils, as well as the features of activities. The underlying assumption is that the whole context of use is
- 13 -
significant for collaborative learning and not only the tools themselves. This description will support the design analysis and discussion that follow in the next unit. System’s Identity Elements
SYSTEMS PURPOSE 1 DIALAB / To teach argument and critical thinking 2 CoVis / To transform science learning to better resemble the authentic practice of science 3 Belvedere / To teach collaborative inquiry 4 Knowledge Forum / To build your community Knowledge 5 COLER / To solve database-modeling problems. 6 C-CHENE / To teach modeling and the concept of energy in physics. 7 BetterBlether / To develop communication skills in unsupervised group discussion. 8 Group Leader Tutor / To promote collaboration skills during the course of problem solving discussions. 9 DEGREE/ To increase the effectiveness of the learning process, by promoting collaboration skills 10 COMET To teach a group of engineers how to work together on software design problems.
Age
Task characteristics
Students
Rigid logic-based dialogue game
High school
Open-ended inquiry in science learning
9-12th grade
Scientific inquiry, investigation of real-world “challenge problems” Creation of a Knowledge building community.
Schools (From 4th grade students), Universities, for everyone Students that have the right level of domain knowledge for using the system. Students (16-17 years old).
Database-modeling. Modeling in physics.
Primary school classrooms
Discussion on any topic
Students
Discussion on problem solving
Students
Discussion
Adults
Object-oriented design problems.
Table 1: Main activities and pupils of ten characteristics collaborative learning systems
DIALAB DIALAB (Moore, 1993) is a synchronous communication tool, designed to teach argument and critical thinking skills using a rigid logic-based dialogue game. Students use the Dialab system in pairs. It has a set of dialogue moves (allowable move type), a set of commitment rules (used to monitor the statements each player has committed to during the dialogue) and a set of dialogue rules (which determine which move types can follow - 14 -
each other). When using the system one of the students starts by choosing a move from the set of sentence openers. Turn-taking is rigidly enforced and a participant can only make one dialogue move per turn. Following a turn, the computer system updates the sender’s commitment store (using the commitment rules) and the move and statement records. The commitment rules keep track of each participant’s commitment store and these are updated following each turn and are visible to both participants. Finally win-lose rules are applied to the collaborative dialogue, identifying situations in which a participant won or lost the game (for example by showing inconsistency in the commitment store).
BELVEDERE Belvedere’s (Suthers and Jones, 1997) core functionality is a shared workspace for constructing “inquiry diagrams” which relate data and hypotheses by evidential relations (consistency and inconsistency) (Figure 1). The software also includes a “chat” facility for unstructured discussions, facilities for integrated use with Web browsers, and two artificial intelligence coaches. The first one provides general advice on the structure of the inquiry diagrams from the standpoint of scientific argumentation. It helps the students to understand the principles of inquiry. The other coach performs various comparisons between the students’ diagrams and an inquiry diagram provided by a subject matter expert. This coach can provide students with feedback concerning correctness, or confront students with new information (found in the expert’s diagram) that challenges students in some way.
- 15 -
Figure 1: Belvedere Inquiry Diagram and Advice
COVIS CoVis, (Learning through Collaborative Visualization) (Pea et al., 1994), explores issues and scaling, diversity and sustainability as they relate to the use of networking technologies, to enable high school students to work in collaboration with remote students, teachers and scientists. Participating students study atmospherics and environmental sciences through inquirybased activities. CoVis provides students with a range of collaboration and communication tools. These include: desktop video teleconferencing; shared software environments for remote, real-time collaboration; access to the resources of the Internet; scientific visualization software; and a special multimedia scientist’s notebook. The notebook (Figure 2) is a medium for students to record their thoughts and actions as they perform scientific inquiry. A notebook consists of a title page with a brief description of the notebook’s purpose, a table of contents and any number of content pages. The table of contents for a notebook displays the notebook’s title, a list of its authors and an overview of its pages including their types, titles and relational structure. Each page in a
- 16 -
notebook has a type, a title and a set of authors. Pages may be authored by individuals, or by a group of people working together at the same time. The types given to pages by their authors provide both descriptions of their contents and of their relationship to other pages.
Figure 2: A Conjecture page of CoVis Notebook. Buttons on the side of the text activates links to other pages.
COLER COLER (Constantino-Conzalez & Suthers, 2001a; 2001b) is a web-based collaborative learning environment in which students can solve database-modelling problems while working synchronously in small groups at a distance. COLER is designed for sessions in which students first solve problems individually (private workspace) and then join into small groups to develop group solutions (public workspace) (Figure 3). The initial problem-solving helps ensure individual participation and provides differences between students’ solutions that form the basis for discussion. When all the students have indicated readiness to work in the group, the shared workspace is activated and they begin to pace components of their solutions in the workspace. Only one student can update the shared workspace at a given time. A panel shows the name of the student who has the control of this area and the students waiting for a turn.
- 17 -
Figure 3: COLER Collaborative Student Interface
The students can get information about Entity-Relationship Modeling while COLER’s personal coach encourages students to discuss and participate during collaborative problem solving. It observes participation in the shared workspace and in chat discussions. Using this information COLER decides whether to give advice.
KNOWLEDGE FORUM Knowledge Forum (Scardamalia & Bereiter, 1994; Hakkarainen & Lipponen, 1998) allows users to create a knowledge-building community. Each community creates their own database in which they can store notes, connect ideas, and “rise-above” previous thinking, based mainly in a asynchronous collaboration mode. Users start with an empty database to which they submit ideas, share information, reorganize the knowledge. They can enter text and graphic notes into the database on any topic they are working on. All users on the network can read the notes and users may build on, or comment on, each other ideas. Authors are notified when comments have been made or when changes in the database have occurred. ‘Knowledge Forum’ makes information accessible with multiple vantage points and multiple entry points. It is also significant that the display of the
- 18 -
community’s work can be organized in multiple and flexible visual representations (Figure 4).
Figure 4: Knowledge Forum Uusers view the knowledge base from multiple perspectives, thus discovering new connections.
C-CHENE C-CHENE (Baker and Lund, 1997) is a CSCL environment for learning modelling and the concept of energy in physics. A pair of two students (16-17 years old) construct their energy chains together in this graphical interface and all of their discussion takes place via specially designed synchronous communication interface. The full screen is divided into two parts from top to bottom, by two buttons for shifting “mode” between “construct” energy chains and “communicate” (Figure 5). The “communicate” screen area contains three windows: one chat-box for each of the students, a structured chat box and a dynamically updated interaction-history trace. The student can observe all actions on the screen of each other, in real time. The interface enforces structured collaboration and turn-taking.
- 19 -
Figure 5: C-CHENE "dedicated" communication interface
DEGREE Degree (Barros and Verdejo, 2000) is an asynchronous CSCL system which characterizes group and individual behaviour in terms of a set of attributes. The system analyzes the state of collaboration using a model of interactions and offers advice intended to increase the effectiveness of the learning process. This is accomplished by requiring users to select the type of contribution (e.g. proposal, question or comment) from a list each time they add to the discussion. The system rates the collaboration between pairs of students along four dimensions: initiative, creativity, elaboration and conformity. These attributes, along with others such as the length of contributions, factor into a fuzzy reference procedure that rates students’ collaboration on a scale from “awful” to “very good”. The advisor elaborates on the attribute values and offers students tips on improving their interaction.
- 20 -
BETTERBLETHER BetterBlether (Robertson, Good and Pain, 1998) is a computer mediated communication environment for use in primary school classrooms. It provides structured support for children developing their communication skills in unsupervised group discussion. The system can be used to discuss any topic. When pupils start a session with BetterBlether, the first question is displayed on the screen to stimulate their ideas and opinions about the topic. When the group feels that has fully covered the question, they have the option of moving on to another questions. The user can send messages to a particular group member or to the group as a whole. The message is constructed from one of the coloured sentence openers buttons followed by additional text. The sentence openers are arranged and coloured by skill category, but names of the skill categories are not stated. The skill categories are not taught explicitly, because some sentence openers could fit into more than one category. All discussion contributions are logged to a text file, which can then be reviewed by the teacher or by the group members themselves. COMET It is a collaborative learning system designed to teach group of engineers how to work together on software design problems (Soller 1999; Soller, Wiebe & Lesgold, 2002). The users use a task window to construct, and can share these diagrams with their group members. The system combines speech acts and sentence openers in a structured interface to help students provide effective help to their peers and encourage them to engage in active learning. Sentence openers are grouped by speech acts types in the interface. The interface also displays logs of the numbers of each type of speech act used in the conversation and the number of contributions from each group member. The idea behind this display is that when students have access to information of this type, they are in better position to
- 21 -
diagnose and overcome problems in the group interactions. Further more, the system is actually designed to advise students, based on Hidden Markov Models that could provide information on appropriate collaboration episodes that influence knowledge sharing (Soller, Wiebe & Lesgold, 2002). GROUP LEADER TUTOR ‘Group Leader Tutor’ (McManus and Aiken, 1995) is intended to promote the collaboration skills identified in Johnson and Johnson (1994) during the course of problem solving discussions between two students. The students send messages to each other by selecting a sentence opener from a menu and then elaborating on this opener with additional text. An intelligent tutoring system offering advice and feedback on the student’s skill is used during the course of the discussion and generates feedback at the end of the discussion. The tutoring system’s suggestions are based on the concept that a conversation can be understood as a series of conversational acts (e.g. Request, Mediate) that correspond to users’ intentions. Unlike other systems (e.g. ‘Coordinator’ by Flores et al., 1988), users are not restricted to using certain acts based on system’s beliefs. Group Leader monitors contributions from the students and compares them to an ideal model of interaction. This might be problematic, because a system, which uses no natural language understanding, is not in an ideal position to criticize students’ discussion skills through analysis of sentence openers.
MAIN TOOLS AND FUNCTIONS TO SUPPORT COLLABORATION The analysis of the previously presented systems as well as the analysis of other developed collaborative systems shows that the central and multiple tools of these systems are related to the support of some high level functions and facilities during collaboration: (A) The means of dialogue and actions,
- 22 -
(B) The functions of workspace awareness, (C) The facilities related to students’ self-regulation or guidance, (D) The facilities related to teachers assistance, (E) The functions related to community level management. The collaborative learning community, up to the present, has devoted the most extensive work, on the main means of action and communication, while the other aspects are under investigation. A) The Means of Dialogue and Actions Before discussing about these means, it would be useful to categorise collaborative learning systems according to the kind of collaborative activity that implicates students, given that the main means of dialogue and actions that students dispose depends on the learning activity itself. Thus, systems can be distinguished in three categories: (a) Action oriented collaborative systems: Some collaborative systems are based on the idea of starting from actions on specific representations, as to express and capture their emerging knowledge and then make this knowledge-representation itself a subject of artifact-centered discourse. This is the case, for instance, of action oriented systems based on disciplinary representations such as the presented C-CHENE, COLER, COMET, or of other developed systems such as, Algebra-JAM (a system for primary school students; Wu et all. 2002), or MODELLINGSPACE (a collaborative system for interdisciplinary modelling in secondary education; Dimitracopoulou & Komis 2003). (b) Text-production oriented systems: A category of collaborative systems invites students to produce in a collaborative or cooperative way a written text or report. For instance, in Knowledge Forum and CoVis, students have to create text-based files presenting their point of view on a topic, or the report of a whole activity. This constitutes the principal activity usually addressed to wide groups that focus on building shared knowledge, into a
- 23 -
community of learners. In this category also belong systems supporting collaborative argumentative writing, such as COSAR (Erkens et all, 2002) (c) Argument-oriented systems: All these systems have in common a linking of different comments or arguments relevant to an issue, usually with categorisation of hyperlinks of their targets with labels such as “argument’, ‘counter-argument’, ‘hypothesis’, ‘example’, etc. constituting argument mapping environments, respecting various representational notations (alongside other existing disciplinary representations). Belvedere belongs to this category, but also a number of other similar systems have been developed such as: Convince Me (Ranney et all, 1995), SenseMaker (Bell, 1997), Representation (Komis, et al. 2002). The discussion on the main means of dialogue and actions, of these three categories of systems will be facilitated by the distinction of the following five involved aspects that designers have to consider: (a) modes of communication, (b) means of dialogue (c) structure of dialogue, (d) structure of actions - coordination protocols
A.1.) Modes of Communication Collaborative learning requests interactions during the learning process for accomplishing the common goals, creating the feeling that collaborators work closely, but also that they belong at the same learning community. The participants in a CSCL system should be able to exchange ideas, to argumentate on their actions or on their points of view, to send and receive messages or materials. There are (a) synchronous collaboration environments which require students and instructors to be consistently connected and in constant attendance, (b) asynchronous collaboration environments which allow parties to communicate in a disconnected fashion and eradicate barriers of time and/or place, (c) Mixted systems: There are systems that support both, as is shown in Table 2. The
- 24 -
production of mixted systems supporting both synchronous and asynchronous collaboration expresses the actual tendency of designers, even if there is always a first dominant mode of interaction, depending of the nature of the collaborative activity. A.2). Means of Dialogue and Discourse Formalisms The main tools that students dispose for communication and dialogue in the frame of collaborative text production oriented or action-oriented systems are: e-mail (for asynchronous communication), chat for free dialogue or structured chat, and sticky notes (an annotation tool); while students who work on argument-oriented collaborative systems, often dispose specific tools according to particular representational notations. (a) Electronic mail (e-mail): can be thought of as asynchronous text-based communication. All students in some projects (e.g. CoVis) had their own electronic mail accounts, having the ability to communicate with members of their collaborative group. (b) CHAT: it is a tool for quasi-synchronous communication. The sequential aspects of talk become rearranged, because of utterances are not necessarily related even if they appear in next positions. In general, there is a delay between the production and the posting of the utterance, and this might pose problems for the participants’ sense-making processes. (c) Structured chat interface, with specific sentence openers: The most suitable approach to structure the dialogue, and specially for problem solving collaborative activities, is sentence openers (e.g. I agree because.., I don’t agree.., Please explain…, etc.). The hypothesis is that providing the tight degree of constraint on typewritten communication can promote more focuses on reflection and the fundamental concepts at stake. Problem solving and interaction tasks need to be more interwoven via computer-supported communication in order to allow the emergence of significant interaction and collaboration forms. The structured interface improves students’ collaborative
- 25 -
argumentation because they are encouraged to use certain discourse acts and problemsolving activities (Quignard & Baker, 1999; Baker & Land 1997). It is critical that the openers enable the widest possible range of communications with respect to the learning task. d) Sticky notes or annotation tool: They constitute a more free form of expression and provide a way for the users to specify with more accuracy the most specific ‘objects’ of discussion’ during an action-oriented activity (Kyza et al. 2002). They have the advantage of facilitating references to parts of the artifact recovering the portion of the discussion that is concerned with a given part. Sticky notes are referred as ‘embedded discourse’, given that embed comments directly on or in the display of the artifact under discussion (Suthers, 1999b).
Some empirical evidence supports this embedded approach of
discourse (Guzdial 1997; Wojahn et al. 1998; Feidas et al 2001), allowing students to have more specific and clear discussion. (e) Representation formalisms on Discourse: The argument-oriented collaborative systems dispose specific representational formalisms, in the case discourse is the central activity of collaboration, (e.g. in case of dialogue-based collaborative critical inquiry). Such tools are: (i) Containers: Statements (e.g. hypotheses or empirical observations) are first reported as boxes in the workspace. (ii) Graphs: Statements are recorded at any time as shapes placed in the workplace. Evidential relations are recorded by linking shapes together, (iii) Matrices: Statements are recorded at any time by creating new columns & rows (respectively). Evidential relations are recorded by placing symbols in empty cells. These representations are intended to enable learners to easily record their deliberations (discourse and/or problem solving processes) and eventually conclusions for subsequent reflection and assessment (Suthers, 1999).
- 26 -
A.3. Structure of dialogue Researchers and designers try to influence the dialogue among collaborators, by a mean or another, especially in the action oriented system, in the aim to promote possible learning effects and/or collaboration quality. In most Computer Supported Collaborative Learning Environments communication during collaboration is text-based, there is not a face-to-face communication and hence there is an increased risk of misinterpretation (Moore, 1993). Hence, it is not sufficient to provide distance learners with a communication channel. In parallel, research has shown that the nature of the discussion varies considerably depending on whether the group is supervised or not: the presence of a teacher seems to encourage group members to justify their opinions and elaborate on their comments (Robertson, Good & Pain, 1998). However, a teacher is likely to dominate the discussion. A computer-mediated communication environment might bridge between supervised and unsupervised group work. One idea was that the environment could provide some support and structure in an otherwise unsupervised group discussion. The structuring of the dialogue in a CSCL system aims at: •
Improve shared understanding by making explicit the (underlying) goal of an utterance.
•
Increase task-oriented behaviour and decrease off-task behaviour (Baker & Lund, 1997), so it would be easier for students to focus on specific parts of the problemsolving process.
•
Improve students’ collaborative argumentation (Veerman, 2000). By using a defined set of discourse acts and sentence openers students can be encouraged into certain discourse acts and problem-solving activities, that promote conceptual change and learning (Baker et al., 2001).
- 27 -
Structuring the dialogues can be done by using (a) sentence openers, or discourse acts, that aim at influencing the content of the dialogue and (b) turn-taking, that aim at influencing the “quality” of the collaboration and the equal participation of all collaborators (see Table 2). (a) Sentence openers: Some of the collaborative environments, like ‘Group Leader Tutor’ (McManus and Aiken, 1995), ‘BetterBlether’ (Robertson et al., 1998), and COMET (Soller 1999), structure the dialogues between the students during problem solving discussions. This structuring is usually based on researchers’ previous work, like ‘Johnson & Johnson’ (Robertson, Good and Pain, 1998) and ‘Wegerif & Mercer’ (Pilkington, 2001). The researchers have identified a series of skills and associated subskills that characterize effective group discussions. There is a one-to-one correspondence between sentence openers and the skills identified. Other systems, like ‘C-CHENE’ (Baker and Lund, 1996), structure the dialogue on the basis of analysis of a corpus of ‘chat-box’ interactions, existing models for information dialogues (Moeschler, 1985; Bunt, 1989) as well as analysis of collaborative problem solving interactions (Baker, 1994). (b) Turn-taking: A few systems apply the “turn-taking” approach, in order to ensure that all the learners will participate. For instance, DIALAB has a set of dialogue moves (allowable move type), a set of commitment rules (used to monitor the statements each player has committed to during the dialogue) and a set of dialogue rules (which determine which move types can follow each other). When using the system one of the students starts by choosing a move from the set of sentence openers. Turn-taking is rigidly enforced and a participant can only make one dialogue move per turn. A.4. Structure of the common workspace and coordination protocols In collaborative learning, a shared final product is in general expected from the participants. Thus, in a CSCL system a common and shared workspace usually exists (see
- 28 -
Table 2), for the learners to have a shared point of reference, better mutual understanding and fulfillment of the shared goal. Evidently, all participants must have access to the shared workspace. As a result, the production of the final product must be self-regulated by the participants themselves or coordinated by the system under specific rules. The systems adopt various ways of regulating or coordinating the access to the shared space (see Table 2). Two main kinds of metaphors have been used for a similar coordination protocol of action-oriented systems: a) The Ask/Take Action metaphor, presented as ‘Action Key’ metaphor or ‘ASK/TAKE PENCIL’ metaphor: where the student that disposes the ‘Action Key’ or the ‘Pencil’
controls the common workspace. It appears to be an appropriate choice, applied in the system COLER (Constantino & Suthers, 2001), and COMET (Soller 2001) that seems to provide clear semantics of actions and temporal roles during activity (Feidas, et al., 2001). The holder of the ‘pencil’ or of the “action-enabling key” is the temporally active partner on the level of actions. Through this key request/ key accept/ key reject protocol the active role can change at any point during collaboration, provided that the ‘non active partner’ asks the pencil or the key and the ‘active partner’ accepts the request. This coordination protocol is usually accompanied by a ‘floor control panel’ providing two buttons to control the workspace ASK/TAKE action and LEAVE action. Additionally, this panel shows the name of the student who has the control of this area and the student(s) waiting for his/her turn. b) Another metaphor, this of Traffic-Light is applied in case of a synchronous text production oriented tool (TC3, Erkens, et all., 2002). The turn taking is regulated by a “Traffic-Light”. One student has the green sign and can write the text, the other has the red sign. The student with the green sign can pass on his turn by clicking on the traffic light. Both lights will turn to yellow and flash, signalling that the turn has been asked for.
- 29 -
- 30 -
Table 2 CSCL systems & the Means and the Structure of Communication
Means of action and communication SYSTEMS / Purpose DIALAB / To teach argument and critical thinking
Mode of communication Synchronous
Action No
Private workspace No
Common workspace No
Structure of Communication (on dialogue and action) Structure of dialogues 1. Sentence openers. 2. Set of dialogue rules (which determines which move type can follow each other). 3. Turn-taking (one dialogue move per turn).
There is not a common workspace.
CoVis / To transform science learning to better resemble the authentic practice of science
Synchronous and asynchronous
Use of Scientific Visualization Environment
Yes
Yes
Belvedere / To teach collaborative inquiry
Synchronous and asynchronous
Construction of inquiry diagrams
Yes
Yes
Knowledge Forum / To build your community Knowledge COLER / To solve databasemodeling problems. C-CHENE / Teach the concept of energy in physics.
Synchronous and asynchronous
Building databases
Yes
Yes
1. Sentence openers. 2. Idea networks.
Synchronous
Construction of ER diagrams
Yes
Yes
No structure, only a chat facility
Action Key
Synchronous
Construction of energy chains
No
Yes
At any time, only one student can ‘act’, (with possibility of interruption).
BetterBlether / To develop communication skills
Synchronous
No
Yes
No
1. Set of communicative acts, grouped according to their function. 2. Strict turn-taking (with possibility of interruption). Sentence openers.
Group Leader Tutor/ promote collaboration skills during the course of
Synchronous
No
Yes
No
Sentence openers.
- 31 -
1. Set of page types that provide both a description of their contents, and of their relationship to other pages. 2. Each page has links to other pages that inspired it, and links from the displayed page to other responses. 3. The link types that a page may have are determined by the type of that page. Chat facility.
Structure of the common workspace
1. Discourse acts. 2. Diagrams. Discourse acts and evidential relations (Each statement has a type and at least one link to another statement). 3. Only one person can work at an item at a time. The author can delete a node.
There is not a common workspace.
Group Leader Tutor/ promote collaboration skills during the course of problem solving DEGREE / effectiveness of the learning process. COMET / how to work together on software design problems.
Synchronous
No
Yes
No
Sentence openers.
Asynchronous
No
No
No
Discourse acts.
Synchronous
Yes OMS software design diagrams
No
Yes
Sentence openers.
- 32 -
There is not a common workspace
B. Workspace Awareness The concept of ‘awareness’ of participants in a collaborative learning environment is a central concept that preoccupies designers and researchers. Generally, during a collaborative activity we can distinguish four types of awareness: social, task, conceptual and workspace awareness. Social awareness (Goldman 1992) is the awareness that students have about the social connections within the group (What should I expect from other members of this group? How will I interact with this group? What role will I take in this group?). Social awareness is interpersonal and perhaps best supported implicitly. For example, audio/video conferencing can create communication opportunities that let people exchange necessary information with each other and negotiate their roles. Task awareness is the awareness of how the task may be completed (What do I know about this topic and the structure of the task? What steps must we take to complete the task? What tools are needed to complete the task? How much time is required?) Concept awareness is the awareness of how a particular activity or piece of knowledge fits into the student’s existing knowledge (How does this task fit into what I already know about the concept? What else do I need to find out about this topic? Do I need to revise any of my current ideas in light of this new information?). Support for both task and concept awareness has been considered in collaborative learning and CSCL research; this support often provides explicit structures that students can use as scaffolds to assist them in organization or help them stay focused on the learning tasks. Additionally, the workspace awareness of each collaborator is important for both synchronous and asynchronous collaboration. The on-line workspace awareness is related to the up to the minute knowledge a student requires about other students interactions with the shared workspace (What are the other members of the group
- 33 -
doing to complete the task? Where are they? What are they doing? What have they already done? How can I help other students complete the project?). This awareness is essential if the students have to work together and learn effectively (Gutwin, Stark & Greenberg, 1995). Workspace awareness is necessary for effective collaborative work, but also plays an integral part in how well an environment creates opportunities for collaborative learning. This is important in collaborative learning for two reasons: First, it reduces the overhead of working together, allowing learners to interact more naturally and more effectively. Second, it enables learners to engage in the practices that allow collaborative learning to occur. So the system must display such information as: (a) who is participating; (b) who is controlling the share workspace; (c) what each student is doing (e.g. typing, waiting for the Action Key, moving the mouse, etc.), (d) what is the contribution to the final object for each student (for instance, it could be shown who has introduced what by the names or a color coding). The first and second functionality are usually fulfilled by a ‘floor control panel’. The third one is usually related with the pointing device, e.g. it is highlighted when dragging and moving of objects occurs by the collaborating partner. This mechanism can reduce the ambiguity observed during turn taking (Feidas, et al., 2001). Another more advanced mechanism is the one that adds a gesturing equivalent: when the mouse of the active collaborator passes over a region, the non active collaborator perceives a highlighting of the corresponding region (Suthers et all, 2003). During asynchronous collaboration, off-line workspace awareness, refers mainly to the product history, thus to the question: what is the contribution of each partner to the current product? So, for instance, on the action-oriented systems, the names of
- 34 -
contributors or coded colored cards, could appear under demand. Additionally, visualization structures of students’ discussion and actions with the aid of a suitable Table 3. CSCL systems & Workspace Awareness Awareness
SYSTEMS / purpose
Online AwarenessLearners’ actions DIALAB / argument and critical thinking CoVis / science learning and authentic practice of science
What You See Is What I See (WYSIWIS)
Belvedere / To teach collaborative inquiry
WYSIWIS
Knowledge Forum / To build your community Knowledge
What You See Is What I See.
COLER / To solve databasemodeling problems.
WYSIWIS Team panel: shows which teammates are already connected. Opinion panel: shows teammates ‘ opinions on a current issue. Floor control panel: shows the name of the student who has the control of the common workspace, and the students waiting for a turn. What You See Is What I See.
C-CHENE/ modeling & energy in physics. BetterBlether / Group Leader Tutor/ DEGREE / COMET/
Offline awarenessHistoric of actions
WYSIWIS
A panel shows which teammates are already connected.
1. Optional entry of the names of the learners that “wrote” each node. 2. The inquiry diagram serves as a record of what the students have done. 1. Register of the names of the learners that “wrote” each node. 2. The nodes can be sorted by date, thread or author. 3. Search functions allows for the bringing together of nodes of interest to the searcher. Register of the names of the learners that “wrote” each contribution.
All actions are added, numbered and time-stamped at the interaction history window.
Offline awarenessHistoric of dialogues Each participant’s statements are visible to both participants. 1. Each page has links to other pages that inspired it. 2. Displays an overview for each notebook’s pages, including their types, titles and relational structure. Chat history
Chat history.
All dialogues are added, numbered and time-stamped at the interaction history window. History of the conversation History of the conversation History of the conversation History of the conversation
WYSIWIS
- 35 -
representation can assist the students’ awareness of other’s actions and opinions. For example a graphical overview of the discussion on-line makes it possible for the students to keep track of the discussion (Nakamura, et al., 1999; Veerman & TreasureJones, 1999). Thus, collaborating learners maintain this awareness by tracking information such as the other learners’ locations in the shared workspace, their actions, as well as their interaction history. In Table 3, the approach of the examined systems is reported. Additionally, it was associated information on historic of actions or dialogues, considering that this information increases the collaborators’ awareness feeling. Most of these systems apply the WYSIWIS protocol, without offering a more sophisticated online awareness. C. Support of Self-regulation or Guidance of Collaborative Learning Interactions. How do systems support collaborators to improve the quality of their collaborative activity? We could classify the collaborative learning systems into two general categories according to who is mainly responsible for managing interactions: the participants themselves or the system? For this purpose, we will use the concepts proposed by Jermann (Jermann et all. 2001), differentiating the existing systems in ‘mirroring’ systems, ‘monitoring’ and ‘advising’ ones. 1.) In the first case the learners are expected to manage the interaction by themselves having been given some kind of information about their own interaction. So, a) There are systems that reflect the actions (mirroring systems). They gather data about the students’ interaction, and display this information to the user without any further elaboration. For example, they display a history of the dialogues, or make students aware of the participants’ actions. The data displayed to the students do not undergo any
- 36 -
processing but directly reflect the action taken on the interface. It is then up to the students to interpret the situation and decide what actions (if any) to take. Belvedere, CoVis, CSILE, C-CHENE, BetterBlether etc. belong to this category. b.) There are systems that model the state of interaction via a set of indicators that are displayed to the users (monitoring systems). These systems often focus on quantitative aspects of interaction, like the number of messages, the number of problem solving actions, using different ways of visualization (Soller, 1999; Jermann, 1999; Simoff, 1999; Ogata, Matsuura & Yano, 2000). There are also systems that interpret the content of the interaction. The interpretation of actions by the system is facilitated through a structured interface. The same information is also either intended to be used later by a coaching agent, or to be used by the system in order to decide when to stop the interaction (Dialab). This kind of information either based on quantitative or on qualitative data, offers a metacognitive information to the students. It is to be noticed that it is up to the user to interpret this information, and then regulate better their activity as well as their collaboration quality. 2.) In the second category, it is the system that is expected to manage the interaction between the users (advising systems). The system is responsible for guiding the students toward effective collaboration and learning. It offers advices and is expected to play a role similar to that of a teacher in a collaborative learning classroom. There are systems that: (a) analyze the groups’ conversation in order to decide the quality of the interaction (DEGREE, Group Leader Tutor), and (b) analyze the actions of the participants in order to decide when to mediate (COLER). Summarizing the above, the most basic level of support that a system might offer involves making the students aware of their own and others’ actions. Actions taken on
- 37 -
shared resources, or those that take place in private areas may not be directly visible to the collaborators, yet it may significantly influence collaboration. Enhancing awareness about such actions may help students maintain a representation of their teammates’ activities, that’s what ‘mirroring systems offer by a simple approach of record and reflection of input data. Table 4 A summary of CSCL systems about Guidance of learning interactions SYSTEMS / purpose DIALAB / To teach argument and critical thinking CoVis / To transform science learning to better resemble the authentic practice of science Belvedere/ To teach collaborative inquiry Knowledge Forum / To build your community Knowledge COLER / To solve databasemodeling problems. C-CHENE/ To teach modeling and the concept of energy in physics. BetterBlether/ To develop communication skills in unsupervised group discussion. Group Leader Tutor/ to promote collaboration skills during the course of problem solving discussions. DEGREE/ To increase the effectiveness of the learning process. COMET / To teach a group of engineers how to work together on software design problems.
Input data Messages
Management of the interaction By the students
Monitoring
Messages
By the students
Mirroring
Dialog, shared and private actions Messages, idea networks
By the students
Mirroring
By the students
Mirroring
Shared and private actions, dialog Messages, shared actions
By the system
Advising
By the students
Mirroring
Messages
By the students
Mirroring
Messages
By the system
Advising
Messages
By the system
Advising
Messages
By the students
Monitoring
- 38 -
Intervention Type
Monitoring and advising systems process this input data to obtain a representation, which is either displayed to the collaborators or used by the system. This representation may be quantitative or qualitative in nature. A quantitative derivation process might entail counting, for instance, the number of dialogue or workspace actions a user has taken. A qualitative derivation process requires taking relational information into account, such as interdependencies between actions or between actions and application context D. Teacher’s Support It has been shown that children benefit from group discussions if the class teacher takes an active part in the group interaction (Harwood, 1995). Harwood observed that children as young as eight working in unsupervised groups were unlikely to use questioning and listening skills during the discussion. Their discussions tended to lack continuity and they experienced problems with group relations. The participation of the teacher prompted group members to elaborate on and justify their opinions. Harwood (quoted in Robertson, Good and Pain, 1998) acknowledges that children need the experience of working in an unsupervised group so that they can develop for themselves the skills of managing group relations. Further analysis of Harwood’s results (Robertson, Good and Pain, 1998) suggests that the presence of the teacher decreases the opportunities for children to take the initiative in the discussion, or ask questions and interact with their peers. So, for young student as well as for older one who embark in collaborative learning activities, teachers are expected to spend time interacting with small groups encouraging and supporting the pupils, while in the more advanced stages of curriculum education, groups of pupils work independently of the teacher. Thus, the teacher’s role moves
- 39 -
progressively from group leader, to facilitator, where s/he provides scaffolding for the discussion but does not take an active role, and finally, takes up the role of observer. But observation of several groups, taking place in parallel, is actually not possible, in any system. Consequently, a number of features would be needed to support teachers: Table 5 . CSCL systems and Teacher’s Support SYSTEMS / purpose
Teacher’s support
DIALAB / CoVis / To transform science learning to better resemble the authentic practice of science Belvedere/ To teach collaborative inquiry Knowledge Forum / COLER / To solve database-modeling problems. C-CHENE / To teach modeling and the concept of energy in physics. BetterBlether / Group Leader Tutor /
There is a record with the history of the whole process, available only to the teacher. Student-constructed diagrams provide the teacher with a basis for assessing students’ understanding of scientific inquiry, as well as of subject matter Knowledge. Documents to describe the chronological sequence of events of the collaborative session in reference to a specific student, and the current state of the environment associated with each event. It also includes the existence of chat contributions, but not the exact words.
All discussion contributions are logged to a text file
DEGREE / 1. History of dialogues. 2. The final shared product.
COMET
1. A record of the discussion that has taken place (which could be provided for reference); 2. Information about the task each learner works on, attendance of every learner’s private workspace, attendance of the public workspace, elaborated comparative information
concerning
various
workspaces
differences); and
- 40 -
(e.g
presenting
interesting
3. an elaborated analysis (historic) of the student’s actions in the shared workspace, that each student has contributed. The above functionalities could allow the teacher to detect when particular children are having difficulties; to identify skills, which seem to be generally weak across a group of children and may also suggest when a group of children are not working well together. This might lead to the identification of skills to be addressed through teaching or prompting in teacher supported discussion, or to giving the whole class further teaching on skills identified as weak. It may also provide material for further analysis of group dynamics by the teacher. The actual systems only offer minimum functionalities that could support teachers, as shown in Table 5.
E. Community Management Tools A wide community of collaborators, working synchronously, asynchronously or both, usually has in its disposal some basic complementary tools, such as: a) document sharing spaces, b) whiteboards and c) group formation tools. a) Shared Whiteboard (roughly the same functionality as Usenet-News) provides a forum for general announcement and questions that may be of interest to all environment users. It is usually an asynchronous text-only medium that allows the user to participate in Internet-wide discussion groups open to anyone with access to ‘news’ reading software and an interest in the topic. News provides the opportunity to participate in broad ranging discussions, or more importantly, to post questions on a topic without directing them to a specific individual. CoVis project used the Macintosh client software News Watcher to read news.
- 41 -
b) Document Sharing Space or Shared Document Repository:
We consider that
document sharing management functionality is important for CSCL environments, although not central. Most of the existing systems for collaborative work offer a space for participants to create folders, upload documents and by defining access rights share them with other members of the community. Concerning the previously examined systems, this kind of functionality is supported only by CoVis and Knowledge Forum, systems that were characterised as ‘text-oriented production systems’. BSCW incorporates a very good management system, while the Fle2 learning environment apply a more object-like interface. The actual tendency has to do with two main features: a) Customisation: where the students could change the background, colours or ‘style’ of the document sharing space, label folders, assign icons, arrange and view objects in their preferred way: as a threaded list or icons, etc. (Kligyte & Leinonen, 2001). By setting the preferences the students could create distinct individual spaces, representing their identities. b) Combined Repositories with Annotation facilities: The students could upload their files to a shared space and add short comments to them (Hoppe & Gabner, 2002). Other members of the group could work on the same files and post their versions with comments to the system. An even more advanced feature was applied in Fle2, where the database of different versions could be displayed as a map of linked objects so that the users could trace back the progress or develop new versions of files for the same task following different threads. c) Group Formation Tools: Formation of a learning group, i.e. the identification of all learners, which ‘belong’ to one specific group, can be done outside the system or can be formed inside the system. The group formation inside the system can happen by chance (e.g. all users working on the same task have common communication facilities), by the
- 42 -
learner itself (assigning himself to a group), by the tutor or the administrator, or finally, automatically by the system. A group formation, by the system consists of three main phases: a) Initiation of group formation, b) identification potential co-learners (the system identifies learners who fulfil the requirements for participating in a group with the initiator), c) negotiation with the participants (the system interact with the potential co-learners in order to determine the actual group). Negotiation is a complex process of balancing the goals and the preferences of all participating individuals and the group. Four different approaches are usually applied, when the system is charged with group formation: i) Domain dependent model group formation: These systems support group formation based on knowledge about the learner with respect to the learning domain, using domain dependent learners model. In an intelligent tutoring system for kinematics (Hoppe 1995), when the student ask for ‘help’, the system displays a list of all potential co-learners, according to the appropriate models of these learners. (ii) Learners goal dependent: In the FITS/CL project (Inaba et all, 2000) at the Osaka University, another approach of group formation is taken. The system detects the appropriate situation of starting a collaborative learning session and sets up a learning goal for the learner. Based on the modelling of learning goals for each learner as well as of the whole group the system negotiates with the agents of all learners. (iii) Collaboration context dependent: Wessner & Pfister (2000) take another approach, without a domain-dependent learner model. Taking into account the collaboration context, the authors use the so-called ‘Intented Points of Collaboration’ in the case of a web-based courses. The course tutor defines at which points of the course a
- 43 -
collaborative activity should happen and specifies the group characteristics as well as some activity features i.e. the group size, the max and/or min duration of the groups collaboration, the type of the learning group, the collaboration learning ‘methods’ (Brainstorming, Off-line Discussion, Pro-con Dispute, Explanation, On-line Discussion, etc) and additional material for each activity. Group formation is based-on the learners general profile and the activity (course). (iv) Performance (solution) dependent: In COLER and CONNECT (Baker et all. 1999) the system proposes a dyad group, on the basis of the individuals solutions in their private space, as to constitute productive collaborative dyads that will negotiate and produce a common problem solution. The objective is usually to put together subjects who manifested semantically interesting differences or conflictual solutions, in a way that may give rise to potentially rich argumentation (based on socio-cognitive conflict theory). Researchers apply various criteria to compare solutions and propose dyads constitution. We could consider that, in different learning situations and contexts, different approaches of group formation could be the most appropriate ones. Factors intervened are: collaborative approach, learning mode (class or individual), learning activity, age, etc. It should be noticed that most of the systems used in schools, have not faced group formation design problem. Thus, in these contexts, group formation is usually defined outside the system, by the teachers or the students themselves.
- 44 -
TRADES-OFFS ON COLLABORATIVE LEARNING SYSTEMS’ DESIGN One central aspect of work in design, that is very commonly posed, or at least thought about, is in terms of trade-offs. A designed thing is, usually, one choice among many possibilities that were considered, and even more possibilities that were never considered. Why is a focus on trade-offs important? Because much of critical discussion around collaborative learning takes an extreme position on one or two dimensions of a design trade-off and over emphasizes it at the cost of acknowledging the most basic point that trade-offs are inevitable in design. What we soon come to see is that we have a long way to go in working on our own design space for considering the ways in which collaboration relates to learning and education. There are no easy or obvious answers. In practice quite different arrangements may be possible, preferable, and even practical. Lets consider some important trade-offs in thinking about design of collaborative environments, as examples of these issues. The main trade offs (that is to be) considered by the designers are related to: (1) The means of dialogue; (2) The coordination of the action and dialogue; (3) The selfregulation or metacognition students’ support and teachers’ support;
(4) The differences between the action based systems or the
‘problem solving oriented’ systems and ‘wide community systems.
1) Trade offs on appropriate means of dialogue Systems either action based or dialogue based, even if they dispose a shared workspace to the collaborators, they provide one or more dialogue tools. These means are considered as crucial not only for collaboration but also for learning. Externalisation through written word during collaborative activities, may have significant effects,
- 45 -
specially for the conceptually rich learning activities (for instance related to science or mathematics). Interactive linguistic exchanges among people play an essential role in the elaboration and perpetuation of scientific concepts (Vygotsky), while the primary use and mechanism of acquisition of these concepts is a result of social interaction (Bakhtin). In designing the means of dialogue in a collaborative learning environment one has to deal with at least the following three trade-offs. (a) The trade off between free and structure dialogue (b) The trade off between parallel and embedded communication tools (c)The trade off between text based and oral dialogue tools (a) The Trade off between Free and Structured Dialogue Let’s consider the case of synchronous collaboration. As presented in the previous units, the trade-off is first of all highlighted by the choice between a free chat interface or a structured dialogue interface. It is to be examined in what conditions and for what task users may need each functionality. It appears that the usage of the free versus structured interface is not independent from the type of content being uttered. Research results show (Baker & Lund, 1997) that pairs, who use the ‘free’ communication mode more than the ‘structured’ one, produce more ‘off-task’ statements than the pairs who prefer the ‘structured’ mode. The free chat interface that allows unstructured, synchronous dialogue, seems to be more appropriate during the initial brainstorming phase of problem solving approach, discussion on problem solving or modelling strategy, eventual decisions of the distribution of the task to different members. Furthermore, the management of a problem solution or of a project elaboration is more often expressed by using the free section, while the task and strategy contributions are
- 46 -
more often expressed by the structured one. The Structured chat interface, with specific sentence openers can promote more focuses on reflection and the fundamental concepts at stake (Baker et al.1997; 2001). For this purpose, it is needed to provide the tight degree of constraint on typewritten communication. A great problem remains: do students use the structured interface or do they have the tendency to continue to work in the free chat? Related to this question are objections that we must have in mind when designing CSCL environments: People believe that if the participants of a collaborative learning situation could choose between a structured communication mode and a ‘free’ communication mode they would definitely choose the latter. Bur, some experiments, (Jermann, 1999; Baker & Lund, 1996), have shown that the structured section of the interface was more frequently used than the free section. Requiring learners to select a sentence opener before typing the remainder of their contribution may tempt them to change the meaning of the contribution to “fit” one of the openers, thus changing the nature of the collaborative interaction. For this reason, it is critical that the openers enable the widest possible range of communications with respect to the learning task (Soller, 1999). The sentence openers are not always used as intended, with the result that the contribution following the opener would not necessarily correspond to the discussion skill represented by the sentence opener (Dillenbourg, 2002). This is something that we must have in mind if we analyze the data manually (like ‘BetterBlether’). In all cases, the interest of designers of dialogue tool for collaboration that has the intention to promote learning is broadening and deepening the space of debate and
- 47 -
produce epistemic interactions (Baker et al., 2001), thus interactions that promote argumentative interactions that promote learning. The underlying assumptions are that argumentative interaction promoted by these structured tools, are seen as interactive mechanism inherent in problem solving learning (Doise & Mugny, 1981). Argumentative interactions particularly those that operate in a conceptual plane, can stimulate reflection on subjective explanatory systems, expressed as arguments and deepen restructuring of the way in which domain is conceptualised (Baker 1999, 2002). Finally, it is to be noted that, except the gains that learners may have through a structured dialogue, this dialogue is also crucial for the benefits for significant metaanalysis of collaborative students; this constitutes another advantage of a structured interface. Another kind of structured dialogue tools to be considered are threaded discussion or other tree structure that may be viewed in a summary form. It is a kind of structure that is created just after each dialogue statement, thus does not intervene to the students reasoning during conversation. It is useful for someone to see the history of dialogue (questions and answers of collaborators.). It may be also used to overview questions that are still open, not answered or not discussed at all. Threaded discussions are also useful for asynchronous dialogue. Another form of structured discourse is the one provided by artifacts’ centred discourse tools or argumentative dialogue tools. This may be disciplinary representations or not that can be used either as a stimulus for reasoning and conversation (e.g. Belevedere, Suthers & Jones 1997) or as a medium through which argumentation dialogue can occur (DREW dialogical reasoning educational web tool, Baker et all. 2003). The dialogue tools currently used are mostly Graphs, Containments, and Matrixes. Different tools
- 48 -
encourage different representation formalisms, and therefore different kinds of interactions and learning. Each formalism manifests a particular representational bias (Utgoff, 1986), expressing certain aspects of one’s knowledge better than others. According to Suthers (1998) representational bias manifests in two major ways: a) By its Constraints: that refers to the limits of expressiveness and constraints on the sequence in which knowledge units can be expressed (Stenning & Oberlander, 1995) and b) By its Salience: that expresses how the representation facilitates processing of certain knowledge units possibly at the expense of others (Larkin & Simon 1987). Representational bias constrain which knowledge can be expressed in the shared context, and make some of that knowledge more salient and hence a likely topic for further discussion. One would expect that while graphs and matrixes are the means of dialogue the concept used and the relations; elaboration could be greater than when containers and threaded discussions are used as dialogue means.
e-mail
Threaded chat
chat
Structured chat
Free Annotations and drawings
Post-it annotations
Specific representational formalisms (matrices, graphs, containers,..)
Free concept mapping tools
Figure 2: Text based tools from unstructured to structured and abstract ones
Figure 2, presents an overview of the main dialogue tools, from the unstructured to the structured and abstract ones. A recent research has explored the differences between students working only with a chat, and students working with a chat and a graph dialogue tool (Baker et all 2003). It appeared that students who had in their disposal,
- 49 -
both a chat and a graph dialogue tools, produced more arguments, and also explored further and deepen segments of dialogue than in the chat tool. While, free verbal interaction due to its intrinsic or strategic indeterminacy is an effective means of negotiating meaning, there is still the hypothesis that this effect would be outweighted by the fact that diagrammatic representations are more determinate, and thus more memorable (Ainsworth, Bibby & Wood 1999). The students would thus express more arguments and would be able to see the ‘gaps’ in their space of debate more easily. The trade off, in terms of design, can be resolved by the simultaneous support of a wide range of dialogue tools offered to the users. We consider that it is important to provide students with multiple tools of dialogue, to assure flexibility of use, in different instances, according to their apparent needs in different phases of problem solving as well as according to the needs derived from the complexity of the task.
b) The Trade off between Parallel and Embedded Representations and tools for dialogue A recent trade off appears between the “parallel tools” and the “embedded tools” for dialogue, specially when users work in ‘action driven systems”. Most of the existing systems, offer the shared artifacts and the discussion tools on entirely separate windows. This seems to lead to hardly conducive discourse about artifacts, even if one can work around this problem by placing the discussion tools next to the artifacts under discussion. Suthers (Suthers, 1999) calls these tools parallel communication tools, defined as tools that do not assure any coordination between the discourse and disciplinary representations. In cases of separate artifacts there is a greater distance
- 50 -
between the object of the discussion and the corresponding dialogue, and hence the cognitive load in processing them. Thus, the question is to reflect on if and how to allow ‘embedded discourse representation’ that embeds comments directly in the display of the artifact under discussion. In informal studies, students appear to prefer to embed their discussion directly in the artifact (as comments) rather than switching between artifact and chat (Suthers, 1999a). Because the discourse always takes place in the context of the artifact, embedded communication tools have the advantages of making easier to refer to parts of the artifact or to recover the portion of the discussion that is concerned with a given part. Some embedded communication tools, designed to establish and curry on a discussion in the context of the visual artifact are the following: a) Annotation tools (sticky notes): That allow embedding comments directly on the display of the artifact under discussion. b) Drawing: e.g. disclosing or indicate a representation or a part of a representation (e.g. diagram) under discussion, c) Highlighting: parts of a diagram under discussion. In reality, this option supports ‘gestural deixis’ (Suthers et all., 2003) enhancing the deistic value of the cursor by making its location more visible. If the user passes the cursor over an object, the object is highlighted in a color (e.g. orange), and if the user deliberately selects an object with the cursor, this object is highlighted in another color (e.g. yellow). All these three design options are metaphors of the actions undertaken by pupils when working alone or in a collaborative mode face to face (collocated). Learning sometimes needs more notations than text, which has tended to dominate mainstream CSCL to date. We all recognize the power of progressively constructing or
- 51 -
disclosing a diagram or annotating an image, to enhance an ongoing commentary (by a tutor or student). A limitation of all virtual shared displays is known to be the absence of expressive gesturing. Some empirical evidence supports the embedded dialogue tools approach. Cuzdial (Guzdial 1997) compared ‘anchored collaboration’ (embedded links to discussion threads within artifacts of interest) to newsgroups in undergraduate classes, finding that longer threads of discussion took place, with anchored collaboration. Wojahn et al, (Wojahn 1998) performed an experimental comparison of “split screen’, ‘interlinear (embedded) and ‘aligned’ (side by side) interfaces for collaborative annotations, and found no significant difference in time to task completion but significantly fewer communications in the split-screen representation, which apparently presents the greatest distance between the artifact and the annotations and hence cognitive load processing them. Additionally, other experimentations (Feidas et al. 2002), show that the existence of embedded comments in the form of sticky notes allow students to have more specific and clear discussion. Some disadvantages are that the record of discourse is fragmented across the artifact, making it more difficult to get sense of the whole discussion or to notice relevant relations between discussions about different parts of the artifact, as well as the possibility that the artifact becomes cluttered with comments. It should be suitable to recover chronological versions of the discourse and perhaps to index the discourse in different ways other than those of artifact components or by chronology. The trade-off between parallel and embedded communication tools can be resolved by conceiving linked dialogue representations tools, thus, providing a logical linkage between them that can be viewed in virtually embedded ways if needed. It should be
- 52 -
also useful to switch between parallel and embedded representations (create a note in one representation and view it in another). This approach could resolve the conflict between the typically linear structures of parallel discourse tools and the contextual indexing of embedded discourse representations.
c) The trade off between written and oral dialogue Although this trade off should be the first to mention, it was left at the end due to the fact that it has appeared only recently. Good technical quality of voice groupware is achieved recently, and so this oral dialogue aspect is not yet explored enough from researchers in a collaborative learning context. We still are at an early stage in understanding its pedagogical niche. Voice groupware constitutes eventually one way to avoid synchronous chat, especially for initial or brainstorming discussions on complex projects. In ‘Luceum’s Project’, applied to Open University, for foreign language students, dialogue via audio is used for practice oral skills (Buckingham Shum et al., 2001). The eventual disadvantages for learning purposes are: (a) The sound and the noice produced in a school class context by the groups of students in the computer lab, (b) These data based on free-natural language cannot be analysed for research or metacognitive purposes. (c) The oral dialogue facilitates the communication, but it also facilitates avoidance of more cognitively demanding activities of argumentation by a written way. Actually, researchers can use a mixed approach, providing tools to have oral communication at least for some phases of the collaborative process (initial brainstorming, final refinement, etc).
- 53 -
2. Trade offs on Actions’ Coordination (a) Restricted collaboration protocols vs free ones During collaborative learning a common final product is expected from the participants, so a shared workspace and a shared point of reference exists. In the case of synchronous collaboration, the question that arises is if the production of the final product must be coordinated or it must be left free. The question is applied for action driven systems as well as for text driven systems. In case of restricted collaboration, two main coordination protocol categories were applied: ‘Turn taking’ and ‘Ask/take action’ control protocol. Turn taking protocol was applied mainly to games, considering that it may be too restricting for usual collaborative problem solving. As far as metaphors (“ask/take pencil”, “action key exchange” and ‘traffic light”) are concerned, they could be considered as similar, and there is not any comparative research that assesses their advantages or disadvantages. Eventually, the most recently applied metaphor of ‘traffic light’ presents a more intense appearance (with the orange colour-signal) of the collaborator that has asked to take the control. An implication of this “key exchange” protocol is that deadlocks can be created in cases when one partner cannot proceed with problem solving and at the same time refuses to pass the key over to the other partner. The advantage seems to be that the protocol maintains clear semantics of actions and roles in the shared activity space. This conclusion seems to be in agreement with the view expressed by researchers of similar environments, (Soller, 2001; Avouris et all., 2001). Coordination protocols were applied in all the early systems, eventually because it was easier to implement. Actually, where both approaches are technical possible, it is to re-
- 54 -
examine
the
necessity
of
a
coordination
protocol
(Dillenbourg,
personal
communication, Mai 2002). Finally, there is an interest currently in examining the need or not of communication protocols in the case of oral dialogue. Is the application of an oral dialogue coordination system needed or should there be a free one, were participants are invited to regulate their oral discussion by a social agreement? In ‘Lyceum project’, using a videoconference system without imposing a control (anyone can speak anytime) adult participants ‘learn’ to take turns and maximise flexibility for different kinds of meeting. In such a case, interactional fluidity will be a useful and important key skill for newcomers to learn. Another approach could require ‘microphones’ either to be ‘passed’ among group members themselves, or by a ‘chairperson-groupleader’.
(b) Rights on partners’ contributions modification and the notion of ownership. In fact, the question of coordination protocols is also related to the concept of “workspace awareness”, and the ‘ownership’ of parts of collaborative construct. What are the rights that each partner has on the contributions of the other partner. Some designers have left it free (e.g. in ‘Modeler Tool’, Koch et al., 2001) without any locked mechanism and others prefer to lock them for all other persons than the object owners (e.g. TeamWave; ‘Representation’, Komis, et al. 2002). In order to answer this question, an experiment was organised, (Feidas, et all. 2002). Two alternative collaboration protocols were used; Groups (A) had no ownership control while groups (B) maintained ownership of introduced objects, so partners were not allowed to modify objects introduced by their peers. In the case of groups (B) every time a partner needed to modify an object of different ownership, a negotiation phase had to be initiated
- 55 -
concerning the purpose of the modification, in order to convince the object owner on the proposed modification. On the contrary, the group A without ownership control, displayed disagreeable instances during collaboration. However, discussing further the above-mentioned results, we could consider that eventually students need a clear indication of the ownership (with the direct or indirect indication of names of owners, in each item of the solution), in order to regulate their activity and avoid this kind of conflict. Instead of locking mechanisms, some designers prefer optimistic concurrent control by supporting awareness: indicating who uses currently which component. This may give the student more freedom (all components can be modified) and fosters teamwork. During tests (Koch, Schlichter & Trondle, 2001) severe conflicts were not observed.
- 56 -
3. Trade-off between self regulation/meta-cognition support and teachers support This is a trade off that actually arises, simultaneously to an increasing research interest on the production of tools and functionalities for students’ support and/or teachers’ support. Let us examine at first, the actual possibilities, tendencies and new requirements for self-regulation student’s support. Self regulation skill is referred to one of the meta-cognitive skills that allows a learner to think about his/her own thinking process, and control it, in order to achieve his/her goal by him/her self (Brown 1987). We consider that the concept of metacognition could scale up to a group of individuals who solve a problem, the group viewed as a single cognitive system, according to the distributed cognition assumptions (Salomon 1995, Pea 1995). Systems that contribute in this direction are not these that reflect interactions, (mirroring systems), but those that monitor the state of interaction, by providing collaborators with literal information or visualizations that can subsequently be used to self diagnose and self regulate interaction. The visualizations typically include a set of indicators that represent the state of interaction, possibly alongside a set of desired values for those indicators. The hypothesis is that visualization’ structures of students’ discussion and actions, through a suitable representation can assist students develop meta-cognitive mental action and subsequently self-regulate their collaborative activity. It is useful to examine briefly three current kinds of already developed tools that present visualizations, all of them based on the students’ actions in various collaborative tasks:
- 57 -
(a) Bar-charts & Pies’ like visualisations in problem solving: Jermann (2002) presents visualization tools called “interaction meters”, that represent the number of contributions related to the discussion and to the implementation of the solution. Individuals are presented a constantly updated visualization of their participation in discussion and task related actions through bar charts that show the number of messages and the number of problem solving actions. The design rationale of interaction meters is that they might give to individuals a better representation of their participation and collaborative work organisation as well as of the role they play in problem solving process. Two types of interaction meters were designed. The first compares individuals by representing their participation side by side as two bars (comparative condition). The second represents participation accumulated across individuals, i.e. one bar chart represents the sum of the each participant’s contribution to discussion and another bar represents the sum of the participant’s problem solving actions (accumulated condition). These assumptions are tested experimentally with a “traffic simulator” task and software. The results suggest that the comparative version of the interaction meters is more helpful than the accumulated one and than the absence of interaction meters. Furthermore, results show that co-occurrence of task and interaction regulation allows quicker solving of the problem thus, better performance concerning the given task. Similarly, Zumbach and all. (2002) are in the process of developing an application for co-constructive tasks with functions for tracking, analysing, and feeding back parameters of collaboration to group members, presented by pie-charts. (b) ‘Nested boxes’ in forum discussions: Simmoff (1999) proposed an interesting way to merge the graphical representation of participation rates and the potential of learning. His system visualizes discussion threads with nested boxes. The thinness of the box
- 58 -
edges represents the number of messages produced in response to the opening message for a particular thread. In an educational environment a thicker box might mean deeper conversation, hence deeper learning. (c) Social Networks in a wide community’s exchanges: Finally, in situations where more than two persons interact, social networks may be used to represent the exchange patterns among participants in a discussion (Nurmela, Lehtinen & Palonen, 2003). A social network typically consists of a network of nodes where each node represents a participant. The thinness of an edge connecting any two nodes represents the amount of the discussion between two participants. It is to be noticed that ‘social networks’ like visualisations are not yet used by students themselves, but mainly by researchers. In general, examining current interaction analysis related to literal or numerical information (statistics), or better, implemented visualization tools, that are intended to function as meta-cognitive tools, we should distinguish that: Information may concern the whole group or each member of the group. Analysis may be based only on actions of collaborators, or only on their dialogues. Analysis may concern only the collaboration quality, or the content of the activity. The analysis may be based on indicators that constitute basic indicators (e.g. participation rates) or higher order indicators (e.g. related to collaboration modes, the quality of the solution). It is to be considered that, for instance, in collaborative problem solving, meta-cognition is not only related to the interaction itself but also to strategic reasoning linked to the task. There is the assumption that regulation of the interaction and regulation of the task are closely related mechanisms and their co-occurrence facilitates coordination. Instead, the existing meta-cognitive tools for collaborative activities are presently rather based
- 59 -
on statistical indicators on participation and collaborators actions or messages than on higher order qualitative indicators. These aspects are further discussed in (Dimitracopoulou et al. 2002; Avouris et al. 2003; Jermann, 2001). The whole question of conception and design of appropriate meta-cognitive tools must be further investigated, by the research community, in relation to: the category of students activity (a game or a high cognitive demanding task, content specific, conceptual or not, etc.), the collaboration mode (e.g. common problem solving or tutoring), the age of pupils (young students, adults or university students), their appropriateness for a small group of collaborators or a large group of collaborators (community). In order to develop effective analysis frameworks and tools for collaborative problem solving analysis, we need to investigate some key questions: •
How to coordinate the analysis of actions and dialogues?
•
What are the most significant data to be logged and coded?
•
How to interrelate collaboration features with problem solving content and process and what abstraction methods do we need to eventually construct computational models?
The combination of the above gives an important number of cases. The most important is that those tools could assist the learning process itself, rather than help students organize only their activity process. Up to the present, researchers have focused more on the student’s self-regulation, while they have neglected teachers. The reasons for that are various; to mention just one, research is mostly taken place in laboratories and not in real school context. Students naturally seek the teachers’ help when they realize that more information is needed to
- 60 -
profitably continue an interaction. Teachers also are socially responsible for students learning. We consider that most of the existing collaboration systems present limitations when used by young students in real school settings. Some of the limitations are attributed to the fact that the teacher who is in charge of several students, fails to interpret the enormous number of complex interactions that can take place simultaneously. A crucial actual question is how could we help them fulfill those responsibilities in computer based collaborative situations? Two main questions raised: •
How can teachers help students?
•
How can teachers be supported by appropriate tools to help students?
There has not been done enough research on the significance of the teachers role during collaborative learning through network, and that the fact that the latter can derive useful knowledge from observing or participating with their students in CSCL environments (Lund & Baker, 1999). Some research has focused on the kind of teachers’ interventions, and not on how we could support teachers to proceed to these interventions and what could be their needs during tutoring or coaching of collaborative students. In order to examine the needs of teachers during synchronous collaboration, and determine corresponding requirements, an experimentation was conducted (Petrou & Dimitracopoulou, 2002). The question was to examine teachers’ behavior during synchronous problem solving with known and current learning activities (and not innovative ones). Teachers have applied two complementary scenarios for teachers’ interventions: a) On-line supervision of group of students collaborated in a synchronous mode.
- 61 -
b) Off-line analysis of preceding intervention: each teacher studied the students’ dialogue and action based interaction, then during the next session (next day or two days later) intervened in order to discuss some concepts, or in order to propose new problems. Their needs concerning the way to support collaborative students were among others the following ones: (a) a way to easily supervise multiple groups of students that collaborate in a synchronous mode, (b) a possible presentation of dialogues linked with the actions in the shared space (c) the history of students actions to appear in the final product, which makes easier to see who has contributed, (d) an appropriate and easier mode to take advantage of the detailed logfiles of students interactions. Consequently, it is needed to design and develop tools or partial functions, such as: (a) supervising tools and facilities, (b) an elaborated and linked history of the whole interaction, (c) tools that produce an automated analysis of students’ interactions, based on the logfile related information. It appears that the most difficult requirement to accomplish is the third one. How to provide a rich variety of analysis output, and assist facilitators or experienced learners? The question of analysis is related to the basic questions stated in a few paragraphs previously while discussing on tools for students. A recent work in progress, inscribed in the direction to produce more powerful analysis results is related to the ‘Object– oriented Collaborative Analysis Framework’ (OCAF), (Dimitracopoulou et al. 2002; Avouris et al. 2003). The value of such framework with its corresponding model is mainly related to its capacity to bring up interesting points of view and thus provide information mainly to teachers relating to the quality of the problem solution and the collaboration modes adopted by the participants. The approach has the following
- 62 -
characteristics: (a) It constitutes a both action and dialogue oriented analysis. (b) It allows a multiple layer analysis, starting from basic indicators (related to basic actions and dialogues exchanges) to higher ones. (c) It produces a model that can be expressed in a literal form, while the production of a diagrammatic one is in progress. (d) It could be supported, not only off-line but also on-line. The diagrammatic form of the OCAF model, could contribute in a supplementary way to the analysis, providing a perceptual view. A teacher that examines and compares two diagrammatic models of solutions can directly distinguish, for instance, solution objects that are not appropriate and were not discussed in a group, or others that were discussed a lot and revised. Such information can support teachers to propose intra-group collaboration in order to discuss specific issues. Teachers could easily identify conflict points, or not appropriate approaches and give advice on topics of the debriefing session internal to the group. They could also recognize semantically significant differences between approaches on problem solving and advise further intra-group discussions. A part from, the meta-cognitive tools for students, and the tools for teachers, under development, some systems (named previously as ‘advisory systems’) leave to the system the control of the interaction, providing help and advising students (such as, COLER, Belvedere, GroupLeader Tutor). Actually, different researchers work mainly in one of these three cases (whether the control is attributed either to student or teacher, or even to the system). The underlying design and research work is in progress and is at a merely premature stage. We consider that all the three approaches are valuable, but it would be much more so, if their could be combined in a single learning environment; that would allow for some
- 63 -
actions of control or self-regulation to be divided among the involved agents (collaborators, teacher, system).
4. The trade-off between ‘problem solving systems’ and ‘Wide Community systems’ Up to the present, most designers and researchers have focused their work on one of the two general kinds of collaborative learning systems: (i) systems that promote collaborative problem solving and work with a small number of collaborators (such as, COLER, C-CHENE, COMET, DEGREE), and (ii) systems that are directly addressed to a wide community, usually aiming at collective knowledge building and understanding (such as CoVis & Knowledge Forum). The design of the first category of systems pay stresses more on the tools for shared action, dialogue and meta-analysis, while the design of the second category focuses on the shared document repository, the structure and the multiple visualisation of the material created from the community, the discussion forums, etc. The first category, use more synchronous communication tools, while the second one is mostly based on asynchronous tools. Nowadays, the trade off between these two general categories does not seem to be relevant anymore. On the one hand, researchers on the community-based systems, have recently recognised that it is worthwhile to incorporate some tools and functionalities for synchronous communication and collaboration (Lethinen, 2002, concerning the new release of KnowldgeForum), to organise their work, clarify some concepts/ideas, enhance social awareness. On the other hand, systems for collaborative problem solving, when used in a school environment, can be enriched in their learning objectives when they support exchanges between students’ class: exchange of materials, of ideas
- 64 -
and difficulties, of reports on an inquiry learning process. These exchanges may prove to be worthwhile for both students and teachers. In this sense, we consider that every collaborative problem solving system needs to be accompanied by a community support system, and therefore incorporate tools and features used by the latter. Thus, repositories, group formation and social awareness functions (providing in a semiautomated way information of all new materials added or modifications and actions) are important features in any environment. Currently, this approach is adopted by two, under development, collaborative problem solving systems, such as MODELLINGSPACE and CoLab.
- 65 -
SYNTHESIS ON DESIGN: FUTURE TRENDS ON TOOLS & FUNCTIONALITIES In order to synthesise the tools presented in the previous analysis, and the main tradeoffs related to tools and functions, we are initially based on two main underpinnings: (i) A consideration of all agents and cognitive systems involved in collaborative learning settings: The agents that seem to be considered in some collaborative environments are often one-dimensional. From one hand, during collaboration the main actor is neither only the individual-member of a collaborative team, nor only the team as a whole. Both of these aspects are important, and equally so the case of the whole community formed of individuals and groups collaborating in various modes. On the other hand, in a learning process, (at least in the frame of primary and secondary education), not only the learners but also teachers are involved (independently of their immediately significant or not role). Learner-centred design approach, dominant during the last decade, although it has influenced, positively the designers, it has presented the following drawbacks in collaborative technology-based learning environments: By focusing in principle on the individual learner, it takes out of the cycle the other agents involved (Dimitracopoulou, 2002) that may form one or more cognitive systems, in the sense of distributed cognition theory (Pea, 1995; Dillenbourg, 1996). Consequently, nowadays, all agents involved in the process are important, and may need to have specific tools in their disposal. Thus, it is needed to consider: (a) the individual, (b) each specific team, (c) the whole learners’ community that is formed, as well as (d) the teacher(s). (ii) A complete view of necessary tools and functions supporting collaborative learning: In the ideal case, each agent and each cognitive system needs some basic tools, according to five general functions to be fulfilled in order to allow and support
- 66 -
collaboration aiming at learning progress. (a) Action and discussion functions: leading to action or text production tools, as well as dialogue tools; (b) Course Management: leading to tools for management of material (e.g. repositories), group formation tools, etc.; (c) Workspace awareness’ functions: leading to functionalities related to the immediate workspace awareness, as well as a larger awareness of all the events happened into a wider learning community; (d) Analysis and meta-analysis tools: that support self-regulation and metacognition for students, as well as teachers’ tools for supervising and analysing collaborative interactions, on-line or off-line; (e) Help and Advising functions: leading to simple help systems or more advanced advising systems for students or even for teachers. According to these two first fundamental considerations, let us examine how a generic collaborative learning system functions, how it processes the whole interaction, what functions it assures and to whom it is addressed (presented schematically in Figure 6). The individual user has in his/her disposal the tools for action and dialogue, to eventually work in a private workspace or to interact and collaborate through a shared workspace. The collaborative learning system, internally collects the data of user action, and of each student interaction with other participants, processes these data, and eventually constructs a model of actions and interactions. Then, this system assures five main functions that are necessary to support collaboration. But, according to the first consideration, at least three simultaneous cycles of process are viewed that correspond to the three main agents’ profile: individual, collaborators and teacher. Thus, in order to fulfil individuals’ needs, and support them as individuals in the collaborative process, the system may advise, offers information (visually or verbally/numerical) based in analysis of its activity, or supports some more basic functions such as assurance of the
- 67 -
workspace awareness. In order to support the group of collaborators, it may produce advices, presents information derived from high level indicators through a meta-analysis of collaborative activity), while it is important to support social awareness. In order to manage their productions, participants have in their disposal specific tools (e.g. repositories). Similarly, addressed to the teachers, help functions may be assured, supervision tools, as well as individual, collaborative or even comparative information is presented based on an analysis or meta-analysis of all interactions.
Data Process
Interaction data collection
Construct Model of Interaction
Action data collection
c
cc Collaborators
¡ System
Individual
y
Action Dialogue Tools
Teacher
(a)
Advising & help Workspace awareness
Activity Analysis Advising & help
Repositories Group formation
Social Awareness functions
Activity & collaboration analysis
Management tools
(b)
Help
(e) Meta analysis Tools
Supervision Tools
(d)
(c)
Figure 6: System processes during collaborative activity assuring functions and tools to the involved agents
- 68 -
The above vision of a generic collaborative learning system is based on two additional underlying assumptions: iii) A vision of a mixed category of collaborative learning systems: Analysing what kind of tools are developed per category of systems, it was derived that there are two general categories of systems: (a) systems that are focused on the collaboration between a small group of learners (e.g. working on problem solving), (b) systems that address themselves, from the beginning, to a wide community of learners. These two categories are actually sufficiently developed, given the specific focus of each kind of environment (problem solving or exchange ideas). It is currently possible to develop systems that are enriched from both of these categories, presenting mixed features. iv) A vision of the control of collaborative process as distributed to all the agents: In our point of view, it would be fruitful to work on the direction of expanding the management of the collaboration to all the agents: ‘individual’, ‘collaborators’, ‘teacher’, ‘system’. This expanded management of collaboration would be possible, according to an approach based on a number of general principles, determining when it is needed to intervene (for instance, the system and/or the teacher), as well as who to intervene by taking specific sub-roles. The current approach is often based on a welldefined desired state (a reference model of ‘appropriate collaboration’, already applied in some advising systems), while the management concerns the system. This approach does not seem to be the most appropriate, given that it is valid only in very specific cases of activities, problems, conditions and students’ profile. Generally, knowledge construction activities via large learners’ community, problem solving, or project based collaborative activities, are open and flexible activities while such as model is quite restrictive.
- 69 -
CONCLUSIONS: RESEARCH AGENDA TOWARDS MORE ADVANCED SYSTEMS USED IN REAL SCHOOL SETTINGS There are many ways to promote collaborative learning: collaboration around the computer, collaboration through gadgets, collaboration through a good combination of various existing tools in the web, etc. In this paper, we have based our analysis and discussion on collaborative systems that have been designed for learning purposes and for a narrow or wide category of learning activities. In all these systems, collaborative learning is viewed as a pedagogical method that can stimulate students to discuss information and problems from different perspectives, to elaborate and refine these in order to re- and co-construct (new) knowledge or to solve the problems. In such situations, externalization, articulation, argumentation, negotiation of multiple perspectives, are considered the main mechanisms that can promote collaborative learning (Dillenbourg & Schneider, 1995; Baker, 1996; Veerman, 2000). These systems have allowed new learning settings and in order to support learning and facilitate collaboration, they have managed to develop new cognitive and metacognitive tools. The evolution of research on design and production of collaborative learning systems had has an effect on the appearance of some significant trade-offs related to: the means of dialogue; the coordination of the action and dialogue; the self-regulation / metacognition support of students and the analysis and meta-analysis tools for teachers; the differences between ‘problem solving oriented systems’ and ‘wide community systems’.
- 70 -
Researchers have already extensively investigated main means of action and dialogue during collaborative activity (especially concerning the synchronous one), and thus, some more specific trade-offs have emerged between: free and structured dialogue, parallel and embedded communication tools, or text based dialogue tools and orally based ones. In parallel, we are expecting to have a positive effect from the current effort on standardisation proposals of collaborative learning systems. Specifically, there is one related standard under elaboration (ISO/IEC JTC1 SC36 WG2). The purpose of this standard is to: (a) establish a basis for collaborative learning environment and provide a format for describing a model of the environment, (b) help developers conform to this standard that promotes collaborative learning, and (c) help users of collaborative learning systems select among several, compatible alternatives of components of the collaborative workspace. The standardisation work is currently divided in three axes: (a) the collaborative workspace: in order to establish a reference model for it, (b) the learner to learner interaction scheme, as to provide a data model and the corresponding language to enable the specification of collaborative learning activities and tasks, and (c) the tools’ agent interaction scheme as to provide a description of the characteristics that must be supported in the communication between agents. Concluding, we consider that the research design agenda of the immediate future focuses on the following axes: i) Accentuation of the effort to produce rich and more appropriate systems: The unification of designers’ effort working on different collaborative systems categories, as well as an illuminated and open vision of all the possible ‘human cognitive systems’
- 71 -
formed during various collaborative modes, could produce richer systems that are more appropriate for various human agents collaborating in various conditions and contexts. ii) Elaboration of powerful analysis methods of collaborative interactions: Researchers are in progress to develop methods, that have the potential to derive rich analysis and meta-analysis results, taking into account a number of aspects: (a) the whole content of the activity with both actions and dialogues, (b) the collaboration modes and quality, (c) the content related to the collaboration quality and vice-versa, (d) each agent’s (individual, group, wide community) needs. What have to be derived from this analysis are basic and high level indicators as well as quantitative and qualitative ones. iii) Development of visualised meta-cognitive tools addressed to Students: The previous analysis methods will serve as precious starting point so as to move on from just mirroring systems to monitoring ones. For this purpose, research has to focus on the investigation of appropriate visualisation modes that could produce metacognitive tools, able to support young students on both learning and collaboration process. iv) Development of visualised tools addressed to Teachers: It is just recently acknowledged that one actual new research direction should be related on how we could take profit from the traces/transcriptions of students in order to facilitate the teachers analysis task (Lund & Baker, 1999; Baker, 2001), allowing them to apply diagnosis and thereafter scaffolding. It is needed to provide appropriate analysis and meta-analysis results with appropriate visualisations that could support teachers to intervene if needed in synchronous or asynchronous mode. v) Production of flexible and negotiable environments that respect the sustainability and reusability of the elaborated work: Lessons learned from technology-
- 72 -
based learning environments in schools suggest that we need to consider the school as a community of practice for doing school; creating systems that allow people to act as they are able and then to amplify, transform, and extend their work to new or additional outcomes. Brown and Duguid (2000) argue that information driven technologies and their implementation needed to be grounded in the social life of the school. Given that most of the schools do not have a long history in the exploitation of these environments, it is important to provide flexible architectures and customisable tools, studying how they work in schools, in different cultural and educational contexts. Designers’ research often concludes after a short period of implementation time, without building in the possibilities that students and teachers can adapt and negotiate tools in their perceivable needs (Baker et al. 2001; Dimitracopoulou, 2001). Additionally, it is also fruitful to assure the sustainability and reusability of the work done, designing interoperable systems, open and easily extensible (Hoppe et al., 2002). (vi) Collaborative learning activities and tasks regarding various collaboration modes: We need to keep always in mind that it is not only the features of the applied technology but especially the way of implementation of the technology which support collaboration (Lehtinen, et all 1999). A crucial parallel research agenda concerns the design of appropriate collaborative learning activities and collaborative modes, for different learning purposes and students’ age levels (Dimitracopoulou & Ioannidou, 2003). The effort to elaborate collaborative scripts semantics is promising and do assist designers’ awareness on a rich range of choices (Dillenbourg, 2002). (vii) Exploration of the new possibilities offered by ubiquitous computing and wireless devices: As technology evolves, continuously open new design and research possibilities. Specifically, ubiquitous computing and handheld computers offer new
- 73 -
physical affordances different from these of traditional computer supported collaborative learning applications, by creating more distributed systems (Roschelle & Pea, 2002). Subsequently, what is needed is the investigation of multiple promising device combinations, new tools and functionalities, as well as new interfaces and their corresponding usability.
ACKNOWLEDGMENTS The preparation of the presented paper is partially supported by the “Network of Excellence on Technologies in Distance Learning”, funded by the Hellenic General Secretariat of Research and Technology (network coordinated by University of the Aegean), as well as by ‘ModellingSpace’ Project, School of Tomorrow, IST-2000-25385, (the project is coordinated by the University of the Aegean, Learning Technology and Educational Engineering Laboratory). Their support is greatly acknowledged.
REFERENCES 1. Abrami P.C. & Bures E.M. (1996). Computer Supported Collaborative Learning and Distance Education, Reviews of lead article. The American Journal of Distance Education 10(20): 37-42. 2. Ainsworth S.E., Bibby, P.A., & Wood, D.J. (1999). Analysing the costs and benefits of mutli-representational learning environments. In M.W. van Someren, P.Reimann, H.P.A. Boshuizen, & T. de Jon (eds). Learning with multiple representations. (pp. 120-134). Oxford: Elsevier Science. 3. Anderson M., Jackson, D., (2000). Computer systems for distributed and distance learning. Journal of Computer Assisted Learning, 16, pp213-228. 4. Aronson, E., Blaney, N., Sikes, J., Stephan, G., & Snapp, M. (1978). The Jigsaw Classroom. Beverly Hills, CA: Sage Publication. 5. Avouris N., Dimitracopoulou A., & Komis V. (in press, 2003). On evaluation of collaborative problem solving: Methodological issues of interaction analysis. Journal of Computers in Human Behaviour (JCHB)., Volume 19, Issue 2 or 3, (March- May 2003), Pergamon.
- 74 -
6. Baker M. & Lund K. (1996). Flexibly structuring the interaction in a CSCL environment. In P. Brna, A. Paiva & J. Self (Eds), Proceedings of the European Conference on Artificial Intelligence in Education. Lisbon, Portugal, Spet. 20-Oc.2, pp. 401-407. 7. Baker , M., Quignard, M., Lund, K. & Sejourne A. (in press, 2003). Computer-supported collaborative learning in the space of debate. In (Ed) U. Hoppe, Computer Support for Collaborative Learning: Designing for Change in Networked Learning Environments, CSCL 2003 congress: 14-18 June 2003, Bergen, Norway. 8. Baker M., de Vries E., Lund K., & Quignard M. (2001). Computer-mediated epistemic interactions for co-constructing scientific notions: Lessons learned form a five-year research program. In P. Dillenbourg, A. Eurelings & Hakkarainnen K. (Eds). Proceedings of European Perspectives on CSCL, Maastricht, March 22-24, 9. Baker M., Hansen T., Joiner R., & Traum D. (1999). The role of grounding in collaborative problem solving tasks. In P. Dillenbourg (Ed) Collaborative-learning: Cognitive and Computational Approaches. pp. 31-64, Advances in Learning and Instruction series, Pergamon, Elsevier 10. Baker M., Lund K., (1997a).Promoting reflective interactions in a computer-supported collaborative learning environment, Journal of Computer Assisted Learning, 13(3),175-193. 11. Baker, M.,J., Lund, K., (1997b). Flexibly structuring the interaction in a CSCL environment, COAST research team, UMR-GRIC, CNRS/University of Lyon. 12. Baker, M., (1994). A Model FOR negotiation in Teaching-Learning Dialogues, Journal of Artificial Intelligence in Education, 5 (2), 199-254. 13. Barros, B., Verdejo, F., M., (2000). Analyzing student interaction processes in order to improve collaboration. The DEGREE approach, International Journal of Artificial Intelligence in Education, (2000), 11, 221-241 14. Bell, P. (1997). Using argument representations to make thinking visible for individuals and groups. In Proc. Computer Supported Collaborative Learning ’97, pp. 10-19, University of Toronto. 15. Bodker, S. (1991). Activity theory and the challenge to systems design. In H.E. Nissen, H.K. Klein & R. Hirschheim (Eds). Information systems research: Contemporary approaches and emergent traditions. Amsterdam, Elsevier. 16. Bratitsis T. & Dimitracopoulou A. (2001). Integrated environments for Distance learning via Networks: present and future. In (Ed) V. Makrakis, Proceedings of Hellenic congress with international participation “ New Technologies in education and Distance Education” Rethymno 8-10 June 2001, Atrapos Editions, pp. 78-96.
- 75 -
17. Brown A. (1987). Metacognition, Executive Control, Self regulation, and other more mysterious mechanisms. In F/ E/ Weinert & R.H. Kluwe. Metacognition, motivation and understanding. Hillsdale, New Jersey, Laurence Erlbaum Associates. 18. BSCW: http://bscw.gmd.de/bscw/ 19. Bunkingham Shum S., Marshall, S., Brier J. & Evans T. (2001) .Lyceum; Internet Voice groupware for Distance Learning. In P. Dillenbourg, A. Eurelings & Hakkarainnen K. (Eds). Proceedings of European Perspectives on CSCL, Maastricht, March 22-24, 20. Bunt H.C. (1989). Information dialogues as communicative action in relation to partner modelling and information processing. In (Eds) MM. Taylor, F. Neel & D.G. Bouwhuis., The Structure of Multimodal Dialogue, pp. 47-74. Elsevier Sciences Publishers, NorthHolland. 21. Clancey, W..J. (1997). Situated Cognition: On human knowledge and computer representations. Cambridge: Cambridge University Press 22. COLAB project (IST/School of Tomorrow): http://colab.edte.utwente.nl/index.html 23. Constantino-Gonzalez, M., Suthers, D., D., (2001a).
Coaching Collaboration by
Comparing Solutions and Tracking Participation, In P. Dillenbourg, A. Eurelings, & K. Hakkarainen (Eds). Proc. of EuroCSCL, Maastricht, NL, pp.324-331. 24.
Constantino-Gonzalez, M., Suthers, D.d., Icaza I.J., (2001b). Coaching Web-based Collaborative Learning based on Problem Solution Differences and Participation, In Artificial Intelligence in Education: AI-ED in the Wired and Wireless Future (Proc. AI&ED 2001), J.D. Moore, C.L. Redfield, & W. Lewis Johnson, Eds., IOS Press, 2001, pp 176-187.
25. Dillenbourg P. (2002). Over-scripting CSCL: The risks of blending collaborative learning with instructional design. In P. A. Kirschner (Ed). Three worlds of CSCL. Can we support CSCL (pp. 61-91). Heerlen, Open Universiteit Nederland. 26. Dillenbourg P. (1999a). What do you mean by collaborative learning ? In P. Dillenbourg (Ed) Collaborative-learning: Cognitive and Computational Approaches. , pp. 1-20, Advances in Learning and Instruction series, Pergamon, Elsevier 27. Dillenbourg P. & Traum (1999b). The long road from a shared screen to a shared understanding, In C. Hoadley & J. Rochelle (Eds). Proceedings of 3rd Conference on Computer Supported Collaborative Learning,, Stanford, December 12-15, 1999. 28. Dillenbourg P., Baker M., Blaye A., & O’Malley C. (1996a). The evolution of research on collaborative learning. In Spada E. & Reiman P. (Eds), Learning Human and Machine: Towards an interdisciplinary learning science, pp. 189-211, Oxford: Elsevier. 29. Dillenbourg P., (1996b). Distributing Cognition over humans and machines. In Vosniadou S., De Corte E., Glaser B. & Mandl H. (Eds), International Perspectives on the
- 76 -
Psychological Foundations of Technology-Based Learning Environments, pp. 165-184, Mahwah, NJ: Lawrence Erlbaum. 30. Dillenbourg, P., Schneider, D., (1995). Collaborative Learning and the Internet, University of Geneva, Switzerland. 31. Dimitracopoulou A. & Ioannidou E. (in press, 2003). Design of distributed collaborative activities for young children related to map use and construction. In (Ed) U. Hoppe, Computer Support for Collaborative Learning: Designing for Change in Networked Learning Environments, CSCL 2003 congress: 14-18 June 2003, Bergen, Norway. 32. Dimitracopoulou A.& Komis V. (in press, 2003). Design Principles for an Open and Wide MODELLINGSPACE for Learning, Modelling & Collaboration in Sciences and Mathematics, In (Ed) C. Constantinou “Computer Based Learning in Sciences, Proceedings of Sixth International Conference CBLIS, 5-10 July, 2003, Nicosia, Cyprus 33. Dimitracopoulou A. Avouris N., Komis B., & Feidas C. (2002). Towards open objectoriented models of collaborative problem solving interaction. In (Eds) P. Jermann, M. Mühlenbrock, A. Soller, Workshop Proceedings, “Designing Computational Models of Collaborative Learning Interaction”, 4th Computer Supported Collaborative Learning Conference, (CSCL-2002), Boulder, Colorado, January 2002, pp. 52-28. 34. Dimitracopoulou A., (2001a). Design issues in Learning Environments for young students: The importance and the limits of general usability criteria. In Avouris & Fakotakis (Eds). Advances in Human-Computer Interaction, Proceedings of PC HCI 2001 Conference on Human-Computer Interaction, pp. 115-120, December 2001, Patras, Greece. 35. Dimitracopoulou A. (2001b). Learning environments and Usability: Appropriateness and complementarity of evaluation methods, Proc. 8th Conference on Informatics, “Towards the Information Society”, Nicosia, November 2001.pp. 545-554. 36. Doise W & Mugny G. (1984). The Social Development of the Intellect. Oxford: Pergamon. 37. Edelson,C., Daniel, Kevin O’ Neill, D., (1994). The CoVis Collaboratory Notebook: Computer Support for Scientific Inquiry, School of Education and Social Policy, Northwestern University. 38. Engestrom Y. (1987). Learning by expanding: An activity theoretical approach to developmental research. Helsinki, Finland: Orienta Konsultit Oy. 39. Erkens, G., Kanselaar G., Prangsma M, & Japsers J. (2002). Using tools and resources in computer supported collaborative writing., In G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, January 7-11 2002., pp.389-398
- 77 -
40. Fidas, C., Komis, V., Avouris., N. (2001). Design of collaboration-support tools for group problem solving, in N. Avouris, N. Fakotakis (ed.), Advances in Human-Computer Interaction, pp. 263-268, Typorama Publ., Patras, Greece, 2001. 41. Fle2 (knowledge building supporting system): http://fle2.uiah.fi 42. Florea, M., A., (1999). An Agent-based Collaborative Learning System, University of Bucharest. 43. Goldman, S.V., (1992). Computer Resources for Supporting Student Conversations about Science Concepts. SIGCUE Outlook, 21 (3), p 4-7. 44. Greeno, J.G., Smith, D.R. & Moore, J.L. (1993). Transfer of situated learning. In D.K. Detterman & R. J. Sternberg (Eds). Transfer on trial: Intelligence, cognition, and instruction (pp.99-167). Norwood, N.J.: Ablex. 45. Gutwin, C., Stark, G. & Greenberg, S., (1995). Support for Workspace Awareness in Educational Groupware, CSCL ’95 Proceedings. 46. Guzdial M. (1997). Information ecology of collaborations in educational settings: Influence of tool. In proceedings of the 2nd International Conference on Computer Supported collaborative learning CSCL., December, 1997, pp.91-100. 47. Hakkarainen, K. & Lipponen, L. (1998). Epistemology of inquiry and computer supported learning. Paper presented at the Annual meeting of the American Educational Research Association, San Diego. 48. Harwood, D., (1995). “The Pedagogy of the World Studies 8-13 Project: the influence of the presence/absence of the teacher upon primary children’s collaborative group work.”, British Educational Research Journal, 21 (5), 587-609. 49. Hollan J., Hutchins E., Kirsh D. (2000). Distributed Cognition: Toward a new foundation for Human Computer Interaction Research, ACM Transactions on Computer-Human Interaction, Vol. 7, No 2, June 2000, pp. 174-196 50. Hoppe H.U. (1995) The use of multiple students modelling to parametrize group learning . In Greer J. (ed) Proceedings of the world Conference on Artificial Intelligence in Education AI-ED 95 Washington. 51. Hoppe U. & Ploetzner R. (1999). Can Analytic Models Support Learning in Groups? In P. Dillenbourg (Ed) Collaborative-learning: Cognitive and Computational Approaches., pp. 147-169, Advances in Learning and Instruction series, Pergamon, Elsevier 52. Hoppe U. & Gabner, K. (2002). Integrating collaborative concept mapping tools with group memory and retrieval functions. In G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, January 7-11 2002, pp. 716-725
- 78 -
53. Hutchins E. (1991). The Social Organization of Distributed Cognition. In Resnick L. B., Levine J. M. Teasley S. D. (eds). Perspectives on socially shared cognition. Washington DC: American Psychological Association. 54. Inaba A., Supnithi T, Ikeda M, Mizoguchi R., & Toyoda J. (2000). How can we form effective collaborative learning groups? In Gauthier G., Frasson C, & VahnLehn K. (Eds) Intelligent Tutoring Systems (Proceedings of the IST 2000) pp. 282-291. Berlin: Springer. 55. ISO/IEC JTC1 SC36 WG2 (standard) http://collab-tech.jtc1sc36.org/ 56. Jerman P., Soller A. & Muhlenbrock M. (2001). From Mirroring to Guiding: A Review of State of the Art Technology for Supporting Collaborative Learning, In P. Dillenbourg, A. Eurelings, & K. Hakkarainen (Eds). Proc. of EuroCSCL, Maastricht, NL, pp.324-331. 57. Jermann P. (2002) Task and Interaction regulation in controlling a traffic simulation, in G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, January 7-11 2002, pp. 601-602 58. Jonassen D. (2000). Revisiting Activity Theory as a Framework for Designing StudentCentered Learning Environments, In D. Jonassen & S. Land (Eds). Theoretical foundations of Learning Environments, LEA. 59. Katz, S. & Lesgold, A., (1993). The role of the tutor in computer-based collaborative learning situations. In S. P. Lajoie ans S. J. Derry (Eds.), Computer as Cognitive Tools, Hillsdale, NJ: Lawrence Erlbaum Associates. 60. Kligyte G. & Leinonen T. (2001). Study of functionality and interfaces of existing CSCL/CSCW systems.
Public Report, ITCOLE project, www. http://www.euro-
cscl.org/site/itcole, November 2002. 61. Koch J., Schlichter J., Trondle P. (2001). Munics: Modeling the flow of information in organisations. P. Dillenbourg, A. Eurelings, & K. Hakkarainen (Eds). Proc. of EuroCSCL, Maastricht, NL, pp.348-355 62. Kolodner, L., J., Nagel, K., (1999). The Design Discussion Area: A Collaborative Learning Tool in Support of Learning from Problem-Solving and Design Activities, Proceedings of the Computer Support for Collaborative Learning (CSCL) 1999 Conference, C. Hoadley & J.Roschelle (Eds) Dec, 12-15, Stanford University, Palo Alto, California. Mahwah, NJ: Lawrence Erlbaum Associates. 63. Komis V., Avouris N., Fidas C., (2002). Computer-supported collaborative concept mapping: Study of synchronous peer interaction, Education and Information Technologies vol.7, 2, pp.169-188, 2002 64. Komis, N. Avouris, C. Fidas (in press, 2003). A study on heterogeneity during real-time collaborative problem solving. In (Ed) U. Hoppe, Computer Support for Collaborative
- 79 -
Learning: Designing for Change in Networked Learning Environments, CSCL 2003 congress: 14-18 June 2003, Bergen, Norway. 65. Koschmann, T. (1999). Computer support for collaboration and learning. Journal of the Learning Sciences., 8, 495-497 66. Kuutti K. (1996) Activity theory as a potential framework for human computer interaction research. In B. A. Nardi (Ed) Context and consciousness: Activity theory and human computer interaction (pp. 17-44) Cambridge, MA: MIT Press. 67. Kyza E., Golan R., Reiser B., & Edelson D. (2002). Reflective Inquiry: Enabling group self regulation in inquiry based sciences using the progresss portofolio tools. In G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, January 7-11 2002, pp. 227-236 68. Lehtinen, E., Hakkarainen, K., Lipponen, L., Rahikainen, M., Muukkonen, H., (1998). Computer Supported Collaborative Learning: A Review, Report, University of Turku, University of Helsinki. 69. Lewis R., (1997). An Activity Theory framework to explore distributed communities. Journal of Computer Assisted Learning, 13 (4), 210-218. 70. Lipponen L., (2002). Exploring Foundations for Computer-Supported Collaborative Learning. In G. Stahl, (Ed), 4th Computer Support for Collaborative Learning: Foundations for a CSCL Community, (CSCL-2002), Boulder, Colorado, pp.72-81, LEA, NJ. USA. 71. McManus, M., M.,& Aiken, M., R., (1996). Teaching collaborative skills with a group leader computer tutor, Education and Information Technologies 1 75-96 (1996). 72. Miyake N. (1986). Constructive interaction and the iterative process of understanding. Cognitive Science, 10, 151-177. 73. ModellingSpace project (IST/School of Tomorrow): www.modellingspace.net 74. Moeschler J. (1992). Theorie Pragmatique, acte de langage et conversation, Cahiers de Linguistique Francaise. 75. Moore, D., (1993). Theory of transactional distance. In D. Keegan (Ed), Theoretical principles of distance education (pp. 22-38). London: Routledge 76. Muhlenbrock M. & Hoppe U. (1999). Computer Supported Interaction Analysis of Group Problem Solving. In C. Hoadley & J. Rochelle (Eds). Proceedings of 3rd Conference on Computer Supported Collaborative Learning, Stanford, December 12-15, 1999. 77. Nakamura, M., Hanamoto, K., and Otsuki, S. (1999). Assistance and Visualization of Discussion for Group Learning, Artificial Intelligence in Education, S.P. Lajoie and M. Vivet (Eds.) IOS Press, 1999.
- 80 -
78. Nurmela K., Lehtinen E. & Palonen T., (in press, 2003). Evaluating CSCL Log files by social network analysis. In (Ed) U. Hoppe, Computer Support for Collaborative Learning: Designing for Change in Networked Learning Environments, CSCL 2003 congress: 14-18 June 2003, Bergen, Norway. 79. Ogata, H., Matsuura, K. and Yano, Y. (2000). Active knowledge Awareness Map: Visualizing learners activities in a Webb based CSCL environment, International Workshop on New Technologies in Collaborative Learning, Tokushima, Japan. 80. Paavola S., Lipponen, L., Hakkarainen K. (2002). Epistemological Foundations for CSCL : A comparison of three modes of innovative knowledge communities. In G. Stahl, (Ed), 4th Computer Support for Collaborative Learning: Foundations for a CSCL Community, (CSCL-2002), Boulder, Colorado, January 2002, pp.24-32, LEA, NJ. USA. 81. Pea, D., R., Edelson, D., and Gomez, L., (1994). The CoVis Collaboratory: High school science learning supported by a broadband educational network with scientific visualization, videoconferencing, and collaborative computing, Northwestern University. 82. Pea, R. (1995) Practices of distributed intelligence and designs for education. In G. Salomon (Ed). Distributed cognitions: Psychological and educational considerations (pp.47-87). Cambridge, England: Cambridge University Press. 83. Petrou A. & Dimitracopoulou A. (in press, 2003). Is synchronous computer mediated collaborative problem solving ‘justified’ only when by distance? Teachers’ point of views and interventions with co-located groups during every day class activities. In (Ed) U. Hoppe, Computer Support for Collaborative Learning: Designing for Change in Networked Learning Environments, CSCL 2003 congress: 14-18 June 2003, Bergen, Norway. 84. Pilkington, R., (2001). Analyzing Educational Dialogue Interaction: Towards Models that Support Learning, International Journal of Artificial Intelligence in Education, 12, 1-7. 85. Ploetzner R., Hoppe H.U., Fehse E., Nolte C., Tewissen F. (1996). Model-based Design of Activity Spaces for Collaborative Problem Solving and Learning. In Brna, P., Paiva, A., & Self, J. (Eds.). Proceedings of the European Conference on Artificial Intelligence in Education, pp. 372-378, Lisbon: Colibri. 86. Puntambekar, S., (1999). An integrated approach to individual and Collaborative learning in a web-based learning environment, C. Hoadley & J. Roschelle (Eds.) Dec 12-15, Stanford University, Pallo Alto, California. Mahwah, NJ: Lawrence Erlbaum Associates. 87. Quignard M. & Baker M. (1999). Favouring
Modellable
Computer–Mediated
Argumentative Dialogue in Collaborative Problem-Solving Situations. In S. Lajoie & M. Vivet, (Eds). Artificial Intelligence in Education. Open Learning Environments, New Computational Approaches to Support Learning, Exploration and Collaboration, p.129-137
- 81 -
88. Ranney , M., Schank, P. & Diehl, C. (1995). Competences versus performance in critical reasoning: Reducing the gap by using Convince Me. Psychology Teaching Review, 4(2). 89. Resnick, L.B., Saljo, R., Pontecorvo C., & Burge, B. (Eds). (1991). Discourse, tools and reasoning; Essays on Situated Learning. Berlin, Germany: Springer. 90. Robertson, J., Good, J., Pain, H., (1998). BetterBlether: The Design and Evaluation of a Discussion Tool for Education, International Journal of Artificial Intelligence in Education, (1998), 9, 219-236. 91. Roschelle, J. and Pea R. (2002). A walk in the WIDE side: How wireless handhleds may change CSCL. G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, January 7-11 2002, pp. 601-602 92. Salomon G. (1995). Distributed Cognitions: Psychological and educational considerations, Cambridge, England: Cambridge University Press. 93. Scardamalia, M & Bereiter C. (1994). Computer support for knowledge-buidling communities. The Journal of the Learning Sciences, 3, 265-283 94. Schwartz D. L. (1995). The emergence of abstract representation in dyad problem solving. The Journal of Learning Sciences, 4 (3), 321-354. 95. Simoff S. (1999). Monitoring and Evaluation in Collaborative Learning Environments. In C. Hoadley & J. Rochelle (Eds). Proceedings of 3rd Conference on Computer Supported Collaborative Learning, Stanford, December 12-15, 1999. 96. Slavin R.E. (1995). Cooperative Learning. Allyn and Bacon, 2nd edition 97. Slavin R.E. (1997). Research on cooperative learning and achievement: A quarter century of research. Paper presented at the annual meeting of pedagogical psychology, Frankfurt. 98. Soller A., Wiebe J., Lesgold A. (2002). A machine learning approach to assessing knowledge sharing during collaborative learning activities. In G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, pp.128-137 99. Soller, L., A., (1999) Supporting Social Interaction in a Intelligent Collaborative Learning System, University of Pittsburgh. 100. Suthers D., & Hundhausen (2001). Learning by constructing collaborative representations: An empirical comparison of three alternatives. In Proceedings of European Conference on Computer-Supported Collaborative learning, Maastricht, The Netherlands. 101. Suthers D., (1999a). Representational Support for Collaborative Inquiry. In Proceedings of International Conference on System Sciences, IEEE, Hawaii. 102. Suthers D., (1999b). Effects of Alternate Representations of Evidential Relations on Collaborative Learning Discourse In C. Hoaley & J. Rochelle (Eds). Proceedings of 3rd Conference on Computer Supported Collaborative Learning,, Stanford, December 12-15.
- 82 -
103. Suthers, D. & Jones, D. (1997). An architecture for intelligent collaborative educational systems. AIED97, the 8th WorldConference on Artificial Intelligence in Education, Kobe, Japan, August 20-22, 1997 104. Suthers, D., D., Toth, E., E., and Weiner, A., (1997). An Integrated Approach to Implementing Collaborative Inquiry in the Classroom, University of Pittsburg, Learning Research and Development Center. 105. Suthers, D., Girardeau, L & Hundhausen C. (in press, 2003). Deistic roles of external representations in face-to-face and online collaboration. In (Ed) U. Hoppe, Computer Support for Collaborative Learning: Designing for Change in Networked Learning Environments, CSCL 2003 congress: 14-18 June 2003, Bergen, Norway. 106. Teamwave (advanced whiteboard): httt://www.stuffincommon.com 107. Utgoff, P. (1986). Shift of bias for inductive concept learning. In R. Michalski, J. Carbonell, T. Mitchell (Eds). Machine Learning: An artificial Intellgience Approach, Volume II. Los Altos, Morgan Kaufmann, pp. 107-148. 108. Veerman, A. L., & Treasure-Jones, T. (1999). Software for problem-solving through collaborative argumentation. In P. Coitier & J.E.B. Andriessen (Eds), Foundations of argumentative text processing (pp. 203-230). Amsterdam: Amsterdam University press. 109. Veerman, A., (2000).
Computer Supported Collaborative Learning Through
Argumentation, Proefschrift Utrecht University, ISBN: 90-393-2393-3. 110. Wojahn P.G., Neuwirth C.M. & Bullock, B. (1998). Effects of interfaces for annotation on communication in a collaborative task. In proceedings of the conference on Human Factors in Computing Systems (CHI ’98). (pp. 456-463). Los Angeles: ACM Press. 111. Wu, A., Farrell R. & Singley M. (2002). Scaffolding Group Learning in a Collaborative Networked Environment, in G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, January 7-11 2002, pp.245-254 112. Zumbach, J., Mulenbrock, Jansen M, Reimann, , Hoppe, U. (2002). Multi-Dimensional Tracking in Virtual learning Teams: An exploratory study. In G. Stahl (ed). Proceedings of Computer Support for Collaborative Learning, CSCL 2002, congress, Colorado, January 711 2002., pp.650-651
- 83 -
